ECONOMIC AND SOCIAL COMMISSION FOR ASIA AND THE PACIFIC

Vol. 21, No. 1
ISSN 0259-238X
April 2006
ECONOMIC AND SOCIAL COMMISSION FOR ASIA AND THE PACIFIC
COVER PHOTOGRAPH
A group of children in rural India pose enthusiastically for a picture.
(Photograph by Wanphen Sreshthaputra)
Child health in India is at the core of the first article published in this issue of
the Asia-Pacific Population Journal, which assesses the potential for reducing child
and maternal mortality through reproductive and child health intervention
programmes in that country.
The second article looks into the readiness, willingness and ability to use
contraception in Bangladesh, the third focuses on Singapore’s family values and their
impact on fertility while the fourth investigates the role of migration in channelling
Bangladeshi labour to countries of East and South-East Asia.
Also don’t miss this issue’s stimulating Viewpoint article on the
thought-provoking headline “Will HIV/AIDS Levels in Asia Reach the Level of
Sub-Saharan Africa?”
Viewpoint
Will HIV/AIDS Levels in Asia
Reach the Level
of Sub-Saharan Africa?
Major reasons that Asia is unlikely to experience
African-level HIV/AIDS epidemics is that there is an ancient concept of
nationhood and class societies with elites willing to undertake national
leadership in emergencies, together with a different sexual culture.
By John C. Caldwell*
The short answer to the question posed by the paper’s title is “no”, although
any disease that kills millions should be combated with all the means available. I
first addressed this question in an editorial in Social Science and Medicine a decade
* Demography and Sociology Program, Research School of Social Sciences, HC Coombs Bldg 9, The
Australian National University, Canberra ACT 0200, Australia. E-mail: [email protected].
Asia-Pacific Population Journal, April 2006
3
ago (Caldwell, 1995) and little has changed since then. That view is supported by
the evidence found in major recent reports upon which this viewpoint is based
(Stanecki, 2004; UNAIDS/WHO, 2004; USAID/US Census Bureau, 2004).
The Asian and Pacific region should not relax. Nevertheless, the comparison
of the region’s present or likely future with that in Eastern and Southern Africa is
absurd and dangerous. It tells us less about the disease than it does about the
psychology and politics of donor international agencies and recipient national
Governments. The picture presented is of a disease sweeping forward from one
area to another as “Spanish” Influenza did in 1918 or the Black Death did in Europe
in the mid-fourteenth century. Certainly, this seemed a plausible scenario in the
1980s after AIDS was first identified in Los Angeles, United States of America in
1981. In the course of that decade we found that it existed in most parts of the
world, much of the spread doubtless taking place during the 1970s, and the pattern
established by the end of the 1980s has, with some intensification, remained in
place. The strongest evidence for the “sweeping forward” thesis of the spread of
AIDS is provided by the failure of high HIV levels to develop in Southern Africa
until the 1990s when the region took over world leadership in the density of
infection. Equivalent examples are hard to find.
One can understand the very different levels of HIV only if it is realized that
the causative retrovirus is in most societies difficult to transmit. Outside the body it
dies very quickly; indeed most of its transmission is from one person to another in
body liquids, usually blood, semen or mother’s milk. Different societies are
characterized by very different chances of this happening. This depends on many
factors, of which the nature of the family, the position of women and attitudes to
sexuality are paramount. This is shown by a level of adult HIV prevalence in
sub-Saharan Africa of 7.4 per cent compared with 0.2 per cent in adjacent North
Africa. Adult prevalence in Southern Africa is over 20 per cent, a level which
means that almost half of deaths are due to AIDS and that life expectancy may be
halved. In comparison, the levels in Japan, China and India are of the order of 0.02
per cent, 0.1 per cent or higher and 0.5 per cent, respectively, not markedly
different from Australia and the United States at 0.15 and 0.6 per cent. What shows
how precise the conditions must be for major epidemic to take hold is the situation
across sub-Saharan Africa itself. Although the societies of Southern and Western
Africa are not very different, adult prevalence levels are over 20 per cent in every
country of the former and below 3 per cent in the majority of countries in the latter.
National figures, however, often disguise the danger in which sections of the
population are placed. HIV levels are above average in most urban populations and
4
Asia-Pacific Population Journal, Vol. 21, No. 1
far above average among most sexually active homosexuals, commercial sex
workers and intravenous-drug users (IDUs). Typically in sub-Saharan Africa
prostitutes are characterized by HIV levels five or more times greater than those of
other women. In Kathmandu in Asia seropositive levels for pregnant women,
prostitutes and intravenous-drug users are 0.2, 17 and 50 per cent, respectively. In
Japan, as in many other parts of the industrialized world, strict use of condoms
keeps HIV levels among commercial sex workers to near zero, but the level among
homosexuals, some of whom are given to risk-taking, is 2.9 per cent.
How, then, is sub-Saharan Africa so prone to HIV infection? Part of the
difference, as Goody (1976) has argued, is that communally owned land, in
contrast to ancient de facto private property in Asia, meant than control of marriage
and hence of female sexuality, even male sexuality, was less rigid than in Asia.
This allowed for relative female freedom for aeons but left women (and men) at
danger of sexually transmitted infections (STIs) in the nineteenth century as
Europeans disrupted African society and to HIV/AIDS towards the end of the
twentieth century. In sub-Saharan Africa, around 57 per cent of those infected with
HIV are female, partly evidence that women are more susceptible to the virus, but
also a sign that the virus is being transmitted in the general community and that
levels are high. Its existence in the general community means that control is
difficult. In Asia 72 per cent of those infected are males, evidence that much of the
transmission is concentrated in brothels or among the IDU or homosexual
populations. Thus there are focal points where the disease may be controlled.
The conditions for a high-level, African-type AIDS epidemic include the
following: (a) a considerable level of sexual activity outside marriage with female
involvement as well as male; (b) some of that activity involving parallel partners;
(c) a significant level of male sexual activity involving female commercial sex
workers; (d) a carelessness of risk among many participants, often as a result of
alcohol consumption; (e) a low level of condom use, often associated with strong
male resistance to use; (f) a low level of health services allowing ulcerating STIs,
which catalyse HIV infection to persist; (g) a fatalism about death arising partly
from earlier very high death rates; and (h) a scepticism about the cause of the
disease owing something to the fact that it is almost a decade from infection to
symptoms (with only a small proportion of the population being tested for HIV).
Even this concordance of events is often insufficient to set in progress a
major HIV epidemic in the general population (i.e., levels of 10 per cent or more of
the adult population infected, as in Southern Africa, Kenya, Uganda in the past and
the Central African Republic). The additional factor, not in itself a sufficient cause
Asia-Pacific Population Journal, April 2006
5
but multiplying the other factors, is whole societies of uncircumcised males: right
down the main AIDS belt from East to Southern Africa the uncircumcising
societies evidence higher often much higher levels of HIV/AIDS.
The situation in sub-Saharan Africa is improved by there being a
relatively low usage of intravenous drugs (which are expensive in a region
which can easily resort to locally produced non-injecting drugs like Kola nuts
and marijuana). It is aggravated by extraordinary inaction on the part of
Governments (Caldwell, 2000). Most of the East and Southern African
countries are facing proportional losses of their 1980 populations comparable
with the losses of the former Union of Soviet Socialist Republics in the Second
World War. In those circumstances one would expect the heads of State to act
like wartime leaders and to forcefully lead their ministers and public services
into the battle to prevent further spread of HIV/AIDS. Except in Uganda, that
has not happened. The leaders regard the battle as hopeless, as something
likely to make them speak of forbidden subjects like sex, as promoting action
which might annoy powerful religious leaders and (in common with most of
their citizenry) as being a lost battle against the biologically based male need to
have more than one partner (a not surprising view in the part of the world with
the highest levels of polygyny). In the new nations, with leaders coming from
specific ethnic groups and being resented by other ethnic groups and in a
situation without an ancient class system, there is often an uncertainty about
giving long-term moral leadership rather than reaping the short-term benefits
of office. The African situation is rendered more difficult by probably the
world’s strongest aversion to condom use and by the dispersal of commercial
and semi-commercial sexual activity so that large easily targeted brothels are
the exception. Certainly, much of female semi-commercial sexual activity
arises from poverty and the need to secure food for the children and protection
for oneself. This picture does not imply that sub-Saharan Africa is an unhappy
place. Far from it, indeed there is an element of happy-go-luckiness that assists
the spread of the disease. The degree of freedom of women is one reason why
migrant Africans tend to fit better into British society than do Bangladeshis or
Pakistanis.
That most of Asia is different and will remain so is suggested by relatively
and persistently low HIV levels. Field research in South India and Bangladesh
showed that young, single rural males periodically made journeys to cities for
business reasons or to visit relatives, and infected no one else. In the rare cases
where a girlfriend was infected, she in turn infected no one else (Caldwell and
6
Asia-Pacific Population Journal, Vol. 21, No. 1
others, 1999). The necessary networks for infection in the general community were
not established. There was almost certainly a markedly lower level of parallel
sexual partners (except in the case of prostitutes) than in sub-Saharan Africa. In
Asian Muslim communities the supervision of females and universal male
circumcision means that HIV levels are everywhere below 0.5 per cent and in the
Arab South-West below 0.1 per cent. Lack of male circumcision among the
majority population of India is probably the main reason its HIV levels are several
times those of neighbouring Pakistan or Bangladesh, although high levels exist
among tribals in two north-eastern hill states.
Major reasons that Asia is unlikely to experience African-level HIV/AIDS
epidemics is that there is an ancient concept of nationhood and class societies with
elites willing to undertake national leadership in emergencies, together with a
different sexual culture. This is true not only of Brahmin or Confucian leaderships
but of their successors in the form of democratic, communist or military leaders.
They are helped by two facts: first, a significant level of commercial sexual activity
taking place in identifiable brothels; second, a lower level among men of
opposition to condom use. Thus the Government of Thailand was able to frighten
many men from going to brothels and frighten most brothel owners into insisting
on condom use. The national seropositive level was stopped from rising above 2
per cent, and there was not a real epidemic in the general population. Even in
Cambodia, where the State has taken a battering and the identity of elites was
savagely attacked, HIV levels appear to have been held to 4 per cent. If any other
Asian State reached half the Cambodian peak level they probably would put into
action programmes similar to the Thai ones.
This is not necessarily true of the whole Asian and Pacific region because
Melanesia differs socially and historically from Asia. Socially and in terms of
sexual activity it is closer to Africa than Asia. The Governments are new and
nations are still being formed. An African-style epidemic could be developing in
Papua New Guinea but an inadequate HIV-testing programme obscures the
situation. Port Moresby and other urban areas may not be large enough to provide
the reservoir of new infections to duplicate the role of African cities in keeping the
epidemic going.
It would be unwise to underestimate the health threat that AIDS (and some
other diseases like tuberculosis and malaria) poses to Asia. There were over a
million new infections in 2004 (compared with over 3 million in sub-Saharan
Africa with less than one fifth of Asia’s population). The projected HIV/AIDS
figures in many mission reports have later had to be scaled down rather than
Asia-Pacific Population Journal, April 2006
7
upward. The frequently proclaimed generalization that there are now more infected
women than men in the world glosses over the fact that this is explained solely by
the predominence of infected women in sub-Saharan Africa (and possibly
Melanesia) and that it has probably always been the case. There is a problem of
crying wolf too often and disheartening those involved in the battle against AIDS.
The Southern African situation is catastrophic but it does not provide a possible
future scenario for Asia. Asia needs not panic programmes but solidly based
long-term public health programmes to minimize permanently disease levels.
8
Asia-Pacific Population Journal, Vol. 21, No. 1
References
Caldwell, B.K., J.C. Caldwell, I. Pieris and P. Caldwell (1999). “Sexual regimes and sexual networking:
the risk of an HIV/AIDS epidemic in Bangladesh”, Social Science and Medicine, vol. 48,
No. 8, pp. 1103-1116.
Caldwell, J.C. (1995). “Understanding the AIDS epidemic and reacting sensibly to it”, Editorial, Social
Science and Medicine, vol. 41, No. 3, pp. 299-302.
________ (2000). “Rethinking the African AIDS epidemic”, Population and Development Review,
vol. 26, No. 1, pp. 117-135.
Goody, J.R. (1976). Production and Reproduction: A Comparative Study of the Domestic Domain
(Cambridge, Cambridge University Press).
Stanecki, K.A. (2004). The AIDS Pandemic in the 21st Century (Washington, D.C., USAID/US Census
Bureau).
UNAIDS/WHO (2004). AIDS Epidemic Update, December 2004 (Geneva, the Joint United Nations
Programme on HIV/AIDS and World Health Organization).
USAID/US Census Bureau (2004). Recent Seroprevalence Levels by Country, September 2004
(Washington, D.C., United States Agency for International Development/US Census
Bureau).
Asia-Pacific Population Journal, April 2006
9
Full text of articles available at
http://unescap.org/esid/psis/population/journal/index.asp
ASIA-PACIFIC POPULATION JOURNAL
Vol. 21, No. 1, April 2006
CONTENTS
Viewpoint
Will HIV/AIDS Levels in Asia Reach the Level of
Sub-Saharan Africa?
3
Major reasons that Asia is unlikely to experience
African-level HIV/AIDS epidemics is that there is an ancient
concept of nationhood and class societies with elites willing
to undertake national leadership in emergencies, together
with a different sexual culture.
This Journal is published
three times a year in English by
the United Nations Economic
and Social Commission for Asia
and the Pacific (ESCAP). The
publication of this Journal is
made possible with financial
support from ESCAP and the
United Nations Population
Fund (UNFPA), through project
number RAS5P203.
The designations employed
and the presentation of the
material in this publication do
not imply the expression of any
opinion whatsoever on the part
of the Secretariat of the United
Nations concerning the legal
status of any country, territory,
city or area, or of its authorities,
or concerning the delimitation
of its frontiers or boundaries.
Mention of firm names and
commercial products does not
imply the endorsement of the
United Nations.
ESCAP is not accountable
for the accuracy or authenticity
of any quotations from sources
identified in this publication.
Questions concerning quoted
materials should be addressed
to the sources directly.
This publication has been
isued without formal editing.
ST/ESCAP/2413
By John C. Caldwell
Abstracts
10
Articles
Potential for Reducing Child and Maternal
13
Mortality through Reproductive and Child Health
Intervention Programmes: An Illustrative Case
Study from India
Some countries may have inadequate technical and financial
resources for improving reproductive and maternal and child
health services. International cooperation both within the
Asian and Pacific region and the wider global community
may be required to meet the challenges of the Millennium
Development Goals at the regional level.
By Minja Kim Choe and Jiajian Chen
Readiness, Willingness and Ability to Use
Contraception in Bangladesh
45
This study has established that most women now want to
control their fertility and consider fertility control as
normatively acceptable, as well as convenient in terms of
availability/accessibility and cost.
By Haider Rashid Mannan and Roderic Beaujot
Singapore’s Family Values: Do They Explain
Low Fertility?
65
Most Singaporeans identified positively with the nuclear
family structure and “standard” family roles. Also,
Singaporeans are generally pro-children. In this context, it is
difficult to support those views that argue that Singaporeans
have become highly individualistic and hedonistic.
By Alexius A. Pereira
International Labour Recruitment: Channelling
Bangladeshi Labour to East and South-East Asia
Given that most host countries in East and South-East Asia
lack a viable alternative, dependence on migrant workers will
be long term even if they choose not to integrate them
permanently into their societies. Any migrant worker policy
has to recognize that such dependence is here to stay.
By Lian Kwen Fee and Md Mizanur Rahman
UNESCAP works towards reducing poverty and managing globalization
85
Editorial Advisory Board:
BENCHA
YODDUMNERN-ATTIG,
Associate
Professor, Institute for Population and Social Research,
Mahidol University, Nakon Pathom, Thailand
BHASSORN LIMANONDA, Director, College of
Population Studies, Chulalongkorn University, Bangkok
CHURNRURTAI KANCHANACHITRA, Director,
Institute for Population and Social Research, Mahidol
University, Nakon Pathom, Thailand
G. GIRIDHAR, UNFPA CST Director for East and
South-East Asia and Representative in Thailand,
Bangkok
GRACE CRUZ, Director, Population Institute,
University of the Philippines, Quezon City, Philippines
JOHN KNODEL, Professor, Population Studies Center,
University of Michigan, Ann Arbor, Michigan, United
States of America
NIBHON DEBAVALYA, former Division Director,
United Nations Economic and Social Commission for
Asia and the Pacific, Bangkok
PHILIP GUEST,
Council, Bangkok
Country
Director,
Population
SHIREEN
JEJEEBHOY,
Senior
Programme
Associate, Population Council, New Delhi
VIPAN PRACHUABMOH, Principal Investigator,
Asian MetaCentre, Singapore
WASIM ZAMAN, UNFPA CST Director for South and
West Asia in Nepal, Kathmandu
United Nations publication
Sales No. E.06.II.F.99
Copyright United Nations 2006
All rights reserved
Manufactured in Thailand
ISBN: 92-1-120472-0
Abstracts
Page
Potential for Reducing Child and Maternal Mortality
through Reproductive and Child Health Intervention
Programmes: An Illustrative Case Study from India
13
According to the assessment made in 2003 by the United Nations Economic
and Social Commission for Asia and the Pacific (ESCAP), UNDP and ADB,
among 47 countries in the ESCAP region with data available, 60 per cent have
already met or are expected to meet Goal 4 of the Millennium Development Goals.
Progression towards Goal 5 has been slower: one third of the countries have
already met the Goal or are expected to meet it. Country-level analysis of available
data suggests that among poor countries, Indonesia and Bangladesh stand out as
making good progresses in meeting Goals 4 and 5. In those countries, reduction of
unwanted and high-risk births through high-level prevalence of contraceptive use
seem to have played very important roles in reducing child and maternal mortality.
India is classified as progressing slowly towards Goal 4 and regressing in
achieving Goal 5 as of 2003. However, 12 out of 17 major states with data,
including some poor states, show declines in under-five mortality exceeding the
level required to achieve the Goal. The present analysis shows that early child
mortality can be reduced substantially in India, beyond the level necessary to meet
Goal 4 through increased utilization of reproductive and child health programmes
even when poverty, women’s education, and community-level sanitary conditions
do not change.
Asia-Pacific Population Journal, April 2006
10
Under-five mortality rate and maternal mortality ratio are highly correlated,
and they share common set of determinants. Thus, the intervention programmes
that would bring about a reduction in under-five mortality rate are likely to reduce
maternal mortality ratio as well.
Readiness, Willingness and Ability to Use Contraception
in Bangladesh
45
This study attempted to measure Easterlin’s notion of motivation or readiness
to control fertility and Coale’s two preconditions of fertility decline-willingness,
and ability. It examined their impacts on the fertility regulating behaviour of
women in view of the rapid fertility decline in Bangladesh. It was observed that
with the exception of women without living children, most women want to control
their reproduction. For most women fertility regulation was found acceptable on
normative and health-related grounds. This is an important finding considering that
it has not been examined before. In addition, for most women family planning
methods are available, accessible and affordable. It is not possible to know from
this study whether and when those conditions have directly played roles in the
Bangladesh fertility decline. However, the study implied that the sociocultural
changes which are favourable to fertility transition have already taken place in
Bangladesh. Regardless of controlling for the background variables, logistic
regression analysis indicated that alike readiness and ability, willingness to
regulate fertility also leads to significantly higher contraceptive use. The three
variables are the principal determinants of contraceptive use and are acting as
intervening variables between most of the background variables and contraceptive
use. However, changes in the background characteristics are key to changes in the
readiness, willingness and ability to use contraception.
Singapore’s Family Values: Do They Explain Low Fertility?
65
Many individuals and policy makers in Singapore believe that society’s
“family values” are being eroded because of processes such as rapid economic
development, industrialization and globalization. They believe that this erosion is
one of the most important reasons why fewer Singaporeans are getting married and
11
Asia-Pacific Population Journal, Vol. 21, No. 1
having fewer or no children, as found in the Singapore Census of Population 2000.
This paper analyses the Singapore-leg of the World Values Survey, conducted in
2002, which contains data on how Singaporeans value family, marriage and
parenthood. It finds that Singaporeans generally still feel that the family is very
important, that marriage is not an outdated institution, and that they would like to
bear several children, if possible. Deeper analysis, however, uncovers some
differences in opinion between “younger” and “older” Singaporeans, which might
suggest that in the future significant value change might take place. This paper
concludes that Singapore’s contemporary family values are unlikely to be the
primary reason behind the declining fertility and marriage rates. It concludes that
Singapore must be facing a “social problem” as there is a sizable gap between
society’s aspirations and the reality.
International Labour Recruitment: Channelling Bangladeshi
Labour to East and South-East Asia
85
International labour recruitment in Asia has been dominated by recruiting
agencies and brokers, who act as intermediaries between workers and foreign
employers. This paper argues that the investigation of temporary labour
migration flows requires examination of the complex infrastructure of
entrepreneurial actors and activities that facilitates labour movement between
two countries and that constitutes the migration industry. Focusing on the
prevailing temporary labour migration programmes, this paper describes the role
of migration institution in the channelling of Bangladeshi labour to the major host
countries of East and South-East Asia. The findings suggest that, despite the
persistent need for migrant workers, some countries in the region, with the
exception of Singapore, pursue non-transparent recruitment policies. This lack of
transparency has mainly contributed to the proliferation of unauthorized
syndicates, and a network of agents, brokers and travel agencies. Bangladeshi
migrant workers risk a large amount of cash to realize a dream overseas and
become vulnerable to victimization.
Asia-Pacific Population Journal, April 2006
12
Articles
Potential for Reducing Child and
Maternal Mortality through
Reproductive and Child Health
Intervention Programmes: An
Illustrative Case Study from India1
Some countries may have inadequate technical and financial resources
for improving reproductive and maternal and child health services.
International cooperation both within the Asian and Pacific region and
the wider global community may be required to meet the challenges
of the Millennium Development Goals at the regional level.
By Minja Kim Choe and Jiajian Chen*
In September 2000, at the United Nations Millennium Summit in New York,
leaders of the world’s Governments signed the Millennium Declaration and
committed themselves to a series of goals and targets that came to be known as the
* Minja Kim Choe, Senior Fellow, Population and Health, Research Program, East-West Center,
United States of America, e-mail: [email protected] and Jiajian Chen, Senior Fellow, East-West
Center, United States.
Asia-Pacific Population Journal, April 2006
13
Millennium Development Goals (MDGs). The Goals include reducing under-five
mortality by two thirds (Goal 4) and reducing maternal mortality ratio by three
quarters (Goal 5) between 1990 and 2015 (IMF, OECD, United Nations and World
Bank Group, 2000). According to the assessment made in 2003 by ESCAP, UNDP
and ADB, among 47 countries in the ESCAP region for which data are available,
one half (24 countries) have already achieved Goal 4 and four additional countries
are expected to achieve the Goal, leaving 19 countries (40 per cent) making slow
progress or regressing. As for Goal 5 (improve maternal health), of the 42 countries
for which data are available, seven have already achieved the Goal and another
seven are expected to achieve it, leaving 28 countries (two thirds) either making
slow progress or regressing. Goal 5 (reduction by three quarters) is more ambitious
than Goal 4 (reduction by two thirds) and it is not surprising that fewer countries
are progressing well towards the first than towards the latter. India is classified as
progressing slowly towards Goal 4 and regressing in achieving Goal 5 as of 2003
(ESCAP, UNDP and ADB, 2005).
In this paper, the authors first examine patterns of major correlates of
under-five mortality rate and maternal mortality ratios, as well as the progress
towards meeting the Goals of reducing under-five mortality rate and maternal
mortality ratio among the countries in the Asian and Pacific region. Doing so,
one hopes to get a better understanding of why some countries are progressing
well towards meeting some of the Goals while some are lagging behind. It is
followed by an in-depth analysis of estimating potential for reducing
under-five mortality through reproductive and child health intervention
programmes including family planning, antenatal care and child
immunization, using India as an illustrative example.
Correlates of under-five mortality rate and
maternal mortality ratio
Recent studies of under-five mortality have identified its key determinants as
poverty, mother’s education, mother’s fertility behaviour (such as age pattern of
fertility, birth spacing and number of births), environmental conditions (such as
source of drinking water and toilet facility), utilization of reproductive and child
health services (such as prenatal care, delivery care and child immunization), and
utilization of health-care services of sick children (Ahmed, Lopez and Inoue, 2000;
Black, Morris and Bryce, 2003; Koenig, Philips, Campbell and D’Souza, 1990;
Miller, Trussell, Pebley and Vaughan, 1992; Mosley and Chen, 1984;
Setty-Venugopal and Upadhyay, 2002; Tulloch, 1999; WHO, 2002; Winikoff, 1983).
Studies on maternal mortality ratio are not as numerous. One of the
difficulties associated with the study of maternal mortality ratio is that it is very
14
Asia-Pacific Population Journal, Vol. 21, No. 1
difficult to collect accurate data especially in countries with high levels of maternal
mortality ratios (UNICEF, UNFPA and WHO, 2004). Limited studies document
that the main causes of maternal mortality are the unexpected complications during
pregnancy, childbirth and other terminations of pregnancy, and just after the
termination of pregnancy, combined with inadequate medical treatment.
Indirectly, knowledge of reproductive health, access to and utilization of
reproductive health care, access to and utilization of medical care, and the
socio-economic and cultural factors associated with knowledge, access and
utilization have been identified as determinants of maternal mortality (UNICEF,
UNFPA and WHO, 2004; Tsui, Wasserheit and Haaga, 1997).
Internationally comparable and accurate time series data on under-five
mortality rates and maternal mortality ratios together with those determinants
would provide an excellent opportunity for in-depth analysis of the causes of
progress or lack of progress on those two Goals. But many countries in the Asian
and Pacific region have limited data available and those are characterized by
varying degrees of accuracy, especially on maternal mortality ratios (ESCAP,
UNDP and ADB, 2005). Therefore, the authors examine the patterns of under-five
mortality rates and maternal mortality ratios at country-level using simple
descriptive statistics and correlation coefficients.
Under-five mortality rates and maternal mortality ratios
in Asian countries: country-level analysis
of the progress towards the Goals
Data
For this part of the analysis, the countries included in East and North-East
Asia are: China; Hong Kong, China; Macao, China; the Democratic People’s
Republic of Korea; Japan; Mongolia and the Republic of Korea. In South-East
Asia they are: Brunei Darussalam, Cambodia, Indonesia, the Lao People’s
Democratic Republic, Malaysia, Myanmar, the Philippines, Singapore, Thailand,
Timor-Leste and Viet Nam, in South and South-West Asia: Afghanistan,
Bangladesh, Bhutan, India, the Islamic Republic of Iran, Maldives, Nepal,
Pakistan, Sri Lanka and Turkey, and in North and Central Asia: Armenia,
Azerbaijan, Georgia, Kazakhstan, Kyrgyzstan, the Russian Federation, Tajikistan,
Turkmenistan and Uzbekistan.2 Data from 37 countries are used for the
country-level analysis. They come from compilations by the United Nations
agencies and the Population Reference Bureau (ESCAP, UNDP and ADB, 2005;
Population Reference Bureau, 2005) but it is to be noted that data are incomplete
for many countries.
Asia-Pacific Population Journal, April 2006
15
Descriptive statistics
Table 1 shows the mean values of indicators of poverty, women’s education,
environment, women’s fertility behaviour, and utilization of reproductive and child
health services for two groups of countries classified according to their progress
towards or distance from Goal 4 of reducing under-five mortality rate. In general,
countries that are not progressing well towards the Goal of reducing under-five
mortality are characterized by high levels of mortality, high levels of poverty, low
levels of education among women, and poor sanitary conditions. In addition, those
countries tend to have high levels of fertility, low levels of contraceptive use, and
early childbearing among women. They are also characterized by a low-level
utilization of reproductive and child health services. But there are some interesting
exceptions. In Bangladesh and Viet Nam, the per capita gross national income (GNI)
is less than US$ 2,500 but the two countries are progressing well towards meeting the
Goal 4 of reducing the child mortality. By contrast, Turkmenistan and Kazakhstan
have relatively high per capita GNI (more than US$ 4,500) but the progression
towards meeting Goal 4 is slow or regressing. Those exceptions suggest that it may
not be necessary to change all determinants of under-five mortality to achieve Goal 4.
Table 2 shows the mean values of the same set of indicators as in table 1 for
two groups of countries classified according to their progress towards or distance
from Goal 5 of improving maternal health. Here, the pattern is less clear than in
table 1. One of the indicators of poverty, the percentage of people with income
less than one-dollar-a-day, for example, is larger in the group of countries that are
progressing well than in the group of countries that are progressing slowly. This
pattern may reflect that one-dollar-a-day may not be a good measure of poverty in
some countries. Countries with small proportions of population with less than
one-dollar-a-day and high maternal mortality include Kazakhstan, Kyrgyzstan
and Viet Nam. It is notable that those countries are also characterized by high
prevalence of induced abortions. In Kazakhstan and Kyrgyzstan, about half of
pregnancies ended with induced abortions in recent years (Academy of Preventive
Medicine, Kazakhstan and Macro International Inc., 1999; Research Institute of
Obstetrics and Pediatrics, Kyrgyz Republic and Macro International Inc., 1998).
The 2002 Viet Nam Demographic and Health Survey reports that about 22 per
cent of pregnancies in the period 1999-2002 were terminated either by menstrual
regulation or induced abortions but considers those to be severely under-reported
(Committee for Population, Family and Children, Viet Nam and ORC Macro,
2003). The high prevalence of induced abortion is likely to be associated with
poor reproductive health of women in general. A closer look at those countries
also reveals that they have large proportions of slum residents among their urban
population. It is likely that urban slum residents have exceptionally high levels of
16
Asia-Pacific Population Journal, Vol. 21, No. 1
maternal mortality. Because the overall level of maternal mortality is generally
low, exceptionally high maternal mortality in a special group can result in high
maternal mortality ratio at national level. Lastly, the authors note that data on
maternal mortality are known to be deficient and inaccurate in many countries
Table 1. Means of selected economic, social, environmental, demographic, and
health-care indicators of countries classifed by their progress towards meeting
Goal 4 (reduce child mortality), countries in the ESCAP region
Mean
among
countries
that have
met or are
expected to
Source meet the
Goal
of data
Under-five mortality rate (deaths per 1,000 live births)
Maternal mortality ratio (deaths per 100,000 live births)
Per capita ppp gross national income, 2004 (US $)
Percentage of population whose income is less than one
dollar a day
Slum population as percentage of urban population
Percentage literate, women aged 15-24
Secondary school enrollment rate, women (per cent)
Percentage with access to safe drinking water (urban)
Percentage with access to safe drinking water (rural)
Percentage with access to improved sanitation (urban)
Percentage with access to improved sanitation (rural)
Total fertility rate
Percentage giving birth in one year, women aged 15-19
Percentage using contraceptives, currently married
women 15-49
Percentage using modern contraceptives, currently
married women 15-49
Percentage immunized against measles, one-year old
children
Percentage births attended by skilled health personnel
Number of countries
Mean among
countries that
are
progressing
slowly
towards the
Goal or
regressing
(a)
(a)
(b)
25
96
11,750
97
374
2,945
(a)
(a)
(c)
(c)
(a)
(a)
(a)
(a)
(c)
(c)
9
22
95
78
95
81
88
65
2.0
3
14
43
85
60
86
61
72
40
3.4
5
(c)
67
44
(c)
51
35
(a)
(a)
90
86
18
80
61
19
Sources: (a) ESCAP, UNDP and ADB (2005). A Future Within Reach: Reshaping Institutions in a
Region of Disparities to Meet the Millennium Development Goals in Asia and the Pacific;
(b) Population Reference Bureau (2005). 2005 World Population Data Sheet;
(c) Population Reference Bureau (2005). Women of Our World.
Asia-Pacific Population Journal, April 2006
17
(UNICEF, UNFPA and WHO, 2004; Tsui, Wasserheit and Haaga, 1997). The
change in maternal mortality ratio, which is the bases for classifying countries into
two groups in the table, is likely to be even less accurate, having to rely on two
possibly inaccurate measures.
Table 2. Means of selected economic, social, environmental, demographic, and
health-care indicators of countries classifed by their progress towards meeting
Goal 5 (improve maternal health), countries in the ESCAP region
Mean among
Mean
countries that
among
are
countries
progressing
that have
slowly
met or are
expected to towards the
Goal or
Source meet the
regressing
Goal
of data
Under-five mortality rate (deaths per 1,000 live births)
Maternal mortality ratio (deaths per 100,000 live births)
Per capita ppp gross national income, 2004 (US $)
Percentage of population whose income is less than one
dollar a day
Slum population as percentage of urban population
Percentage literate, women aged 15-24
Secondary school enrollment rate, women (per cent)
Percentage with access to safe drinking water (urban)
Percentage with access to safe drinking water (rural)
Percentage with access to improved sanitation (urban)
Percentage with access to improved sanitation (rural)
Total fertility rate
Percentage giving birth in one year, women aged 15-19
Percentage using contraceptives, currently married
women 15-49
Percentage using modern contraceptives, currently married
women 15-49
Percentage immunized against measles, one-year old
children
Percentage births attended by skilled health personnel
Number of countries
(a)
56
70
(a)
(b)
184
11,769
289
4,579
(a)
15
10
(a)
(c)
(c)
(a)
(a)
(a)
(a)
(c)
(c)
29
91
68
91
72
80
51
2.5
4
36
91
71
89
68
77
50
2.9
4
(c)
58
53
(c)
48
37
(a)
86
84
(a)
72
16
73
21
Sources: (a) ESCAP, UNDP and ADB (2005). A Future Within Reach: Reshaping Institutions in a
Region of Disparities to Meet the Millennium Development Goals in Asia and the Pacific;
(b) Population Reference Bureau (2005). 2005 World Population Data Sheet;
(c) Population Reference Bureau (2005). Women of Our World.
18
Asia-Pacific Population Journal, Vol. 21, No. 1
Analysis of correlations
In table 3, the correlation coefficients of four dependent variables are
examined (under-five mortality rate, maternal mortality ratio, progress towards
the Goal of reducing under-five mortality, and progress towards the Goal of
improving maternal health) with the determinants of under-five and maternal
mortality. In order to maximize the use of available data, the correlation
coefficients are computed one pair at a time.
Under-five mortality rate is correlated with the indicators of income, women’s
education, national level sanitary conditions in urban and rural parts, fertility
behaviour, and utilization of maternal and child health programmes in a manner
consistent with previous findings. Poverty, low levels of education among women,
poor sanitary conditions, high levels of fertility, high levels of teenage fertility, low
levels of contraceptive use, and low levels of utilization of reproductive and child
health services are associated with high levels of under-five mortality rate. The
magnitudes of the correlation coefficients are high for all correlates and they are all
statistically significant (p<0.05). Similarly, maternal mortality ratio has high
correlations with all the factors examined, in the same direction as the correlation
coefficient with under-five mortality rate, and they are all statistically significant. The
correlation between under-five mortality rate and maternal mortality ratio (not in the
table) is very high (0.82) and statistically significant.
The last two columns in table 3 show correlation coefficients between
whether countries have already met the Goal or are expected to meet it (coded as 1)
or not (coded as 0) and the potential determinants. Most of the correlation
coefficients between the indicator of progress on Goal 4 (reduce child mortality)
and the potential determinants are large and statistically significant.3
By contrast, most of the correlation coefficients between the indicators of
progress on Goal 5 (improve maternal health) and the potential determinants are
small and statistically insignificant. Only one of the determinants, per-capita GNI
has statistically significant correlation with whether the country is progressing well
towards reducing maternal mortality ratio or not. The weak relationship between
the indicator of progress to Goal 5 and the determinants of maternal mortality may
be owing to the measurement problems of the maternal mortality ratio. As
discussed, the statistics on maternal mortality ratio are often inaccurate, and the
progress on maternal mortality, which involves measurements at two or more time
points are much more likely to be inaccurate than the single measure.
When countries are grouped by their level of per capita income, most
countries with ppp (adjusted for purchasing power parity) per-capita Gross
National Income under US$ 4,000 in 2004 are progressing slowly or regressing in
Asia-Pacific Population Journal, April 2006
19
Table 3. Correlation coefficients of selected economic, social, environmental,
demographic and health-care indicators with under-five mortality rate, maternal
mortality ratio, and progress to Goals of reducing them,
countries in the ESCAP region
Progress
Progress
Maternal
Under-five
mortality rate mortality ratio towards Goal towards Goal
of reducing of improving
(deaths per
(deaths per
maternal
child
100,000 live
1,000 live
health
mortality
births)
births
Per capita ppp gross national
income, 2004 (US$)
Percentage of population
whose income is less than
one dollar a day
Slum population as percentage
of urban population
Percentage literate, women
aged 15-24
Secondary school enrollment
rate, women (per cent)
Percentage with access to safe
drinking water (urban)
Percentage with access to safe
drinking water (rural)
Percentage with access to
improved sanitation (urban)
Percentage with access to
improved sanitation (rural)
Total fertility rate
Percentage giving birth in one
year, women aged 15-19
Percentage using
contraceptives, currently
married women 15-49
Percentage using modern
contraceptives, currently
married women 15-49
Percentage immunized against
measles, one-year old children
Percentage births attended by
skilled health personnel
-0.61*
-0.42*
0.50*
0.41*
0.54*
0.74*
-0.20
0.20
0.56*
0.63*
-0.35*
-0.11
-0.43*
-0.80*
0.28
0.00
-0.67*
-0.84*
0.38*
-0.05
-0.79*
-0.86*
0.30*
0.05
-0.72*
-0.56*
0.46*
0.09
-0.76*
-0.74*
0.45*
0.10
-0.67*
0.79*
-0.59*
0.81*
0.50*
-0.53*
0.03
-0.14
0.58*
0.73*
-0.37*
-0.01
-0.79*
-0.73*
0.62*
0.13
-0.59*
-0.51*
0.44*
0.30
-0.59*
-0.74*
0.31
0.07
-0.66*
-0.77*
0.40*
-0.02
Sources: (a) ESCAP, UNDP and ADB (2005). A Future Within Reach: Reshaping Institutions in a
Region of Disparities to Meet the Millennium Development Goals in Asia and the Pacific;
(b) Population Reference Bureau (2005). 2005 World Population Data Sheet;
(c) Population Reference Bureau (2005). Women of Our World.
Note:
* indicates p<0.05.
20
Asia-Pacific Population Journal, Vol. 21, No. 1
both under-five mortality (Goal 4) and maternal mortality (Goal 5), while countries
with ppp per-capita Gross National Income of US$ 4,000 or over in 2004 are
progressing well. However, it is notable that among countries with low levels of
income, Indonesia and Bangladesh are progressing well towards the two Goals
(see appendix tables A1 and A2). A closer look reveals that the characteristics that
separate those countries from other low-income countries are high level of
contraceptive use and low level of fertility. Viet Nam also has low-income, high
level of contraceptive use, low level of fertility. Yet it is progressing well towards
Goal 4. However, despite this remarkable progress on this latest Goal, distance
from Goal 5 is increasing. As speculated earlier, progress towards Goal 5 may be
lagging in Viet Nam owing to the high prevalence of unsafe induced abortion.
Those exceptions suggest that high contraceptive prevalence and low level of
fertility can reduce child mortality by decreasing high-risk births and unwanted
births (Setty-Venugopal and Upadhyay, 2002). This can be achieved even when
economic and development conditions are not favourable to low level of child
mortality and despite the fact that reduction of maternal mortality may be hindered
if the rate of unsafe induced abortion is high.
An attempt was made to conduct country-level multivariate statistical
analysis of progress towards Goals 4 and 5 such as fitting regression models in
order to estimate the “net effects” of each determinant, controlling for the effects of
other determinants. But because many countries have incomplete data (less than 20
countries have complete data available) the estimates become quite unstable.
Obviously, in-depth statistical analysis leading to the estimation of the “net
effects” of the determinants of under-five mortality rate and maternal mortality
ratio or their progresses requires more comprehensive data sources.
Summary of country-level analysis
In summary, country-level analyses show that levels of under-five mortality
and maternal mortality are very highly correlated. Poverty, low level of education
among women, poor sanitary conditions, high level of fertility, high level of
teenage fertility, low level of contraceptive use, and low level of utilization of
reproductive and child health services are associated with high level of under-five
mortality rate and maternal mortality ratio. The cases of Indonesia and Bangladesh
suggest that reduction of under-five mortality rate and maternal mortality ratio can
be achieved by altering some determinants through intervention programmes
aimed at reducing unwanted and high-risk births.
Although most of the determinants of under-five mortality rate and maternal
mortality ratio are correlated, it is likely that reduction of under-five mortality rate
and maternal mortality ratio can be achieved by altering some determinants
Asia-Pacific Population Journal, April 2006
21
through increased levels of utilization of reproductive and child health
programmes. Estimating the potential contribution of intervention programmes,
one needs to have better measures of the “net effects” of the determinants. Such
estimates can be computed from multivariate statistical models as illustrated in the
following case study of India.
Estimating potential for reducing early childhood mortality:
an illustrative case study of India
The case of India
India offers an excellent opportunity for an illustrative analysis to study
determinants of under-five mortality and potential for its reduction for at least three
reasons. The first reason is related to the level of under-five mortality. At national
level, under-five mortality is moderately high and India is progressing slowly
towards meeting the corresponding Goal. According to the assessment made by the
United Nations agencies, India’s under-five mortality in 2003 was 87 per 1,000
live births. This corresponds to the medial level mortality among the 19 countries
in the ESCAP region that are making slow progress towards or regressing on Goal
4 (ESCAP, UNDP and ADB, 2005). However, India is a large and complex
country and there has been large variations at state level in both the level of
mortality and the rate of reduction in mortality in recent years as shown below. The
second reason is related to the state-level variations in factors associated with
under-five mortality. In India, the state government is largely responsible for
implementing reproductive and child health programmes and the utilization of
those programmes vary greatly among states. Similarly, other conditions affecting
under-five mortality such as the level of poverty, sanitary conditions and mother’s
education, vary also greatly among states as shown below. Indeed, the variations
among states of India in terms of under-five mortality and the major determinants
resemble much the cross-national variations among countries observed in the
ESCAP region. Third, the National Family Health Surveys (NFHS) conducted in
the 1990s offer an excellent data source for rigorous statistical analysis. The data
availability is especially important because many of the countries that are making
slow progress towards Goal 4 do not have high quality data that would allow
in-depth analysis.
Under-five mortality in India
In India, the under-five mortality rate was 123 in 1990. Two thirds reduction
to meet Goal 4 means reaching an under-five mortality rate of 41 by 2015. In 2003,
this same rate was estimated at 87, which means that it was reduced by 29 per cent
in 13 years. In order to meet the Goal of reducing the rate by two thirds in 25 years,
22
Asia-Pacific Population Journal, Vol. 21, No. 1
the under-five mortality should have fallen by 35 per cent in 13 years.4 In
summary, in more than half of the time required to achieve the Goal, less than half
of the necessary reduction has been achieved.
Table 4. Child mortality during 1994-1998 and percentage decline in
child mortality during the periods between 1988-1992 and 1994-1998
State/Union Territory
Child
mortality in
1994-1998
Percentage decline Population in
1991 census
in child mortality
(1,000s)
in six years
Decline >=16 per cent
Kerala
Himachal Pradesh
Assam
Delhi
West Bengal
Tamil Nadu
Haryana
Orissa
Karnataka
Gujarat
Bihar
Maharashtra
18.8
42.4
89.5
55.4
67.6
63.3
76.8
104.0
70.0
85.1
105.1
58.1
41
39
37
33
32
27
22
20
20
18
18
17
29,099
5,171
22,414
9,421
68,078
55,859
16,464
31,660
44,977
41,310
86,374
78,937
Decline < 16 per cent
Uttar Pradesh
Andhra Pradesh
Madhya Pradesh
Punjab
Rajasthan
123.0
85.5
137.6
72.1
114.9
13
6
-6
-6
-12
139,112
66,506
66,181
20,282
44,006
Sources: IIPS (1995, p. 221) and IIPS and ORC Macro (2000, p. 194) for child mortality; Office of the
Registrar General (2005) for population.
Note: Small states in the north-eastern region are not included in the table.
According to the estimates based on National Family Health Surveys of
1992-1993 (NFHS-1) and 1998-1999 (NFHS-2), all India experienced a 13 per
cent reduction in under-five mortality from 109.3 in 1992-1993 to 94.9 in
1998-1999, falling somewhat short of the amount of decline required to achieve
Goal 4 (a 16 per cent decline in six years is required) and at about the same rate as
estimated by ESCAP, UNDP and ADB in 2005. However, 12 out of 17 major states
with data show decline in under-five mortality by more than 16 per cent, the level
required to achieve the Goal, as shown in table 4.5 The statistics from Kerala are
Asia-Pacific Population Journal, April 2006
23
most impressive. In this south-western state, not only did the under-five mortality
decline impressively by 41 per cent, it had also reached the level of developed
countries by 1998-1999. This achievement is remarkable considering that the per
capita Gross State Product in Kerala was only Rs. 16,0296 in 1998-1999 (current
prices), far below some other states such as Delhi (34,332 rupees), Goa (40,248
rupees), Punjab (21,194 rupees), and Maharashtra (20,148 rupees) (Ministry of
Statistics and Programme Implementation, 2005). Another noteworthy fact is that
the states that have achieved sufficient amount of reduction in under-five mortality
have varying levels of mortality. Under-five mortality in some states are under 50
per 1,000 live births in the 1994-1998 period, but it is between 50 and 100 in a
number of states, while in Orissa and Bihar under-five mortality is over 100. States
making slow progress or regressing on this Goal also have wide ranging levels of
under-five mortality, from 72 to 138. From this pattern one can conclude that the
slow progress in reducing under-five mortality at the national level is due to slow
progress in some large states which have high levels of under-five mortality such as
Uttar Pradesh, Madhya Pradesh and Rajasthan.
India’s health programmes
The Government of India’s effort to strengthen maternal and child health
services began during the First and Second Five-Year Plans (1951-1956 and
1956-1961) under the Ministry of Health, and continued with the Minimum Needs
Programmes initiated during the Fifth Five-Year Plan (1974-1979). The primary
objective of the effort was to provide basic public health services to vulnerable
groups of pregnant women, lactating mothers and pre-school children. In
1992-1993, the Child Survival and Safe Motherhood Programme continued the
process of integration by bringing together several key child survival interventions
with safe motherhood and family planning activities.
More recently, efforts to improve maternal and child health have been
enhanced by the activities implemented by the Family Welfare Programme
(Ministry of Health and Family Welfare, 1992). Special schemes included the
programme of Oral Rehydration Therapy and the development of Regional
Institutes of Maternal and Child Health in states where infant mortality rates are
high. The Universal Immunization Programme and the Maternal and Child Health
Supplemental Programme within the Post-Partum Programme were also
implemented (IIPS, 1995).
In 1996, the integrated Reproductive and Child Health Programme was
launched incorporating safe motherhood and child health services (IIPS and ORC
Macro, 2000). This new programme seeks to integrate maternal health, child
health, and fertility regulation interventions with reproductive health programmes
24
Asia-Pacific Population Journal, Vol. 21, No. 1
for both women and men. Important elements of reproductive health programmes
include: (a) provision of antenatal care including at least three antenatal visits and
two doses of tetanus toxoid vaccine; (b) encouragement of institutional deliveries
or home deliveries assisted by trained health personnel; (c) provision of postnatal
care; and (d) identification and management of reproductive tract and sexually
transmitted infections.
State-level variations in socio-economic, demographic and health indicators
As discussed in the previous section, the health and welfare programmes in
India are operated jointly by the federal and state governments yet the state
governments play preponderant roles in their implementation. Partly because of
this institutional arrangement and owing also to differing historical and cultural
backgrounds, Indian states vary greatly in socio-economic, demographic, and
health conditions as seen in table 5.
Economic conditions are generally good in the northern states (except
Rajasthan) where less than 10 per cent of households were classified as having low
standard of living index according to the NFHS-2 survey. Eastern states are
economically least advanced with about half of the households classified as having
low standard of living index. Women’s illiteracy varies greatly according to states,
from a low of 10 per cent in Mizoram and 13 per cent in Kerala to a high of 77 per
cent in Bihar and 76 per cent in Rajasthan. Other states in central and eastern
regions show high levels of illiteracy among women.
In terms of fertility, Goa, Kerala, Himachal Pradesh, Punjab and Tamil Nadu
show very low fertility and relatively late ages at marriage. In some states in the
southern region such as Karnataka and Andhra Pradesh, total fertility rate is low
but early marriage is quite common. By contrast, in some states in the north-eastern
region such as Mizoram, Nagaland and Manipur, fertility is high but early marriage
is uncommon. In many large states such as Rajasthan, Madhya Pradesh, Uttar
Pradesh, and Bihar, fertility is high and early marriage is widespread.
In most of the states in the north, west and south, the prevalence of antenatal
tetanus toxoid vaccine (two or more times)7 is quite high, reaching 95 per cent in
Tamil Nadu and more than 85 per cent in Delhi, Punjab, Goa and Kerala. States in
central, eastern and north-eastern areas lag behind substantially. Among big states,
it is only 51 per cent in Uttar Pradesh and 52 per cent in Rajasthan. Similar patterns
are observed in childhood immunizations,8 but the gap is wider ranging from more
than 80 per cent in Himachal Pradesh, Goa, Kerala and Tamil Nadu to less than 25
per cent prevalence in Rajasthan, Madhya Pradesh, Uttar Pradesh, Bihar,
Arunachal Pradesh, Assam, Meghalaya and Nagaland.
Asia-Pacific Population Journal, April 2006
25
Table 5. Selected socio-economic, demographic and health indicators,
major states/unions territories of India, 1998-1999
State/Union
Territory
North
Delhi
Haryana
Himachal Pradesh
Jammu & Kashmir
Punjab
Rajasthan
Central
Madhya Pradesh
Uttar Pradesh
East
Bihar
Orissa
West Bengal
North-east
Arunachal Pradesh
Assam
Manipur
Meghalaya
Mizoram
Nagaland
Sikkim
West
Goa
Gujarat
Maharashtra
South
Andhra Pradesh
Karnataka
Kerala
Tamil Nadu
Percentage
Percentage of women Percentage
Percentage Percentage
of women receiving
illiterate,
of houseTotal aged 25-49 two antena- of one-year
everholds with
fertility married tal tetanus olds with
married
low stanfull immunrate
toxoid
before age
women
dard of livization (b)
innoculat18
ing index aged 15-49
ions (a)
3
10
8
9
4
23
29
55
36
70
39
76
2.4
2.9
2.1
2.7
2.2
3.8
38
60
38
48
23
82
85
80
66
78
90
52
70
63
83
57
72
17
31
29
69
70
3.3
4.0
79
80
55
51
22
21
53
51
46
77
60
50
3.5
2.5
2.3
84
58
62
58
74
82
11
44
44
23
39
34
44
13
25
13
53
54
43
38
10
40
49
2.5
2.3
3.0
4.6
2.9
3.8
2.8
40
49
21
35
13
24
35
46
52
64
31
38
51
53
21
17
42
14
60
14
47
15
22
23
29
50
45
1.8
2.7
2.5
15
54
65
86
73
75
83
53
78
36
30
15
34
64
55
13
48
2.3
2.1
2.0
2.2
80
61
27
42
82
75
86
95
59
60
80
89
Source: IIPS and ORC Macro (2000).
Notes: (a) Among women who gave births during the five-year period before the survey, for the last
and next-to-last births.
(b) Among one-year olds at the time of survey.
26
Asia-Pacific Population Journal, Vol. 21, No. 1
Analytical strategy
Using data from the National Family Health Survey (NFHS-2) conducted in
1998-1999, the authors first estimated a statistical model to estimate effects of
major determinants of early childhood mortality (ages 0-24 months). The
estimated statistical models of early childhood was then used for simulation by
altering the values of major determinants. Most of the simulation consisted of
estimating the reduction of mortality under the hypothetical situation specified by
the conditions in Kerala, the state with the lowest level of infant and child
mortality in all India. Some additional simulations, involving additional changes
in fertility behaviour and reductions of sex differentials in childhood mortality
were also undertaken. The simulation exercises showed the extent of potential
reduction in early childhood mortality through reproductive and child health
intervention programmes.
Data
To estimate the effects of major determinants of early childhood mortality,
data from National Family Health Survey (NFHS-2) conducted in 1998-1999
were used. The survey is based on a nationally representative sample of 91,196
households and on all ever-married women aged 15-49 within the households
(89,199 women). The sampling fraction varies from state to state, in order to
assure that the sample size in each state is large enough to provide statistically
meaningful estimates (IIPS and ORC Macro, 2000).
In each state, the rural sample was selected in two stages. The first stage
consists of the selection of the primary sampling units (PSUs), which are villages,
with probability proportional to population size (PPS). The second stage consists
of the random selection of about 30 households within each PSU. In urban areas, a
three-stage procedure was followed. In the first stage, wards were selected with
PPS sampling. In the next stage, one census enumeration block (CEB) was
randomly selected from each sampling ward. In the final stage, households were
randomly selected within each sample CEB. On average, 30 households were
targeted for selection in each selected enumeration area (IIPS and ORC Macro,
2000).
Each household is assigned a state-level weight and an all-India weight.
Weights are needed to correct for over-sampling of some groups and
under-sampling of others. The all-India weights take into account variability in
sampling fractions among the states. The calculation of weights takes into account
non-responses as well. In the analysis in the paper, all-India weights, normalized
so that the sum of weights is equal to the number of observations in the sample, are
used.
Asia-Pacific Population Journal, April 2006
27
Statistical model for early childhood mortality
The survey collected full birth histories of women, including information on
birth order, date of birth, sex, whether child was alive at the time of survey and for
the children who have died, age at death. For the last and next-to-last births
occurring since January 1995, the survey collected information on antenatal care
including antenatal visits, tetanus toxoid vaccine, place of delivery, and birth
attendance. For the surviving children among those, information on feeding,
children’s vaccination, morbidity, and care of sick children were also collected.
The authors first created the children file consisting of one record for each
child born to women in the survey during the five-year period preceding the
survey, extracting information from birth histories of women. Then the child
record was used as the unit of analysis. The analysis was limited to children born
during the five-year period before the survey because many of the household-level
and community-level factors examined were measured at the time of the survey
and would not have been accurate for children born long before the survey. Some
child-specific factors such as year of birth, survival status of the child, age at death,
sex, birth order, mother’s age at birth, previous birth interval, sex combination of
surviving older siblings at the time of birth, whether any of the older siblings have
died at the time of birth of the index child, and previous birth interval are computed
and added to the child record.
Figure 1. Monthly probability of dying estimated from NFHS-2
among children born during the five-year period before survey
Deaths per 10,000
500
450
400
350
300
250
200
150
100
50
0
0
3
6
9
12
15
18
21
24
Month
28
Asia-Pacific Population Journal, Vol. 21, No. 1
To estimate the effects of factors on early childhood mortality, the hazard
model (Cox model) for the month-specific probability of dying during the first 25
months (months 0 through 24) of life was used. Hazard model is chosen so that the
analysis does not have to be limited to children who have been observed for the full
25 months since birth. The hazard model allows for the inclusion of children who
were born less than 25 months before the survey (censored cases). The mortality
during first 25 months rather than a more usual 24 months was used in order to
include reported deaths at age 24 months. Age at death data in India tend to heap at
6, 12, 18 and 24 months as shown in figure 1. Limiting analysis to deaths at 0-23
months, one would be excluding some deaths that occurred at age 19-23 months
but reported as occurring at age 24 months. According to the NFHS-2 data, about
90 per cent of under-five mortality in India takes place at ages 0-24 months. Thus,
the present analysis should include most of the under-five mortality.
Hazard model requires two dependent variables: survival status and exposure
time. The survival status is measured at the end of age 24 months or at the time of
survey for those who were born less than 25 months before the survey. The
exposure time is the age at death for those who died by the end of age 24 months,
and age at survey for those who were surviving.
Three sets of factors were examined: household/mother-level factors,
child-level factors, and community-level factors. Our choice of factors is
determined on the basis of major literature on under-five mortality in general, and
in India (Black, Morris and Bryce, 2003; Mosley and Chen, 1984; Pandey and
others, 1998). The household/mother-level factors include the standard of living
index of the household created by IIPS (low or not), and the level of mother’s
education (none, primary and more).
The child-level factors include year of birth, sex of child, and mother’s age at
birth for all children. For children of birth order two or higher sex combination of
surviving older siblings was also included, along with preceding birth interval,
birth order, and the dummy variable indicating whether any of the older siblings
have died before the birth of the index child. All child-level factors are created to
reflect the condition at the time of birth of the index child.
The first community-level factor is the urban-rural designation of the
community. Two community-level factors were also included indicating sanitary
conditions: percentage of households with piped water and percentages of
households with toilet facility. Lastly, two factors indicating the level of utilization
of reproductive and child health programmes were included: percentages of
women in the community who received at least two tetanus toxoid vaccines among
those who gave birth during the five-year period before the survey,9 and
Asia-Pacific Population Journal, April 2006
29
percentages of one-year olds in the community who received full childhood
vaccinations.10 It would have been ideal to include the utilization of reproductive
and child health programmes as child-level variables as well. Unfortunately,
however, this information was collected only for last or next-to-last births, and for
childhood vaccination, only for the children who were still surviving at the time of
survey, making it impossible to treat them as potential factors of early childhood
mortality.
By contrast, using those measures at community-level have benefits.
Estimating the effects of the utilization of reproductive and child health
programmes at the community-level is sensible because many of the programmes
are related to the control of infectious diseases at the community level. The primary
sampling units (PSUs) are identified as communities. Usually, one PSU in rural
area consists of one village and one PSU in urban area consists of one census
enumeration block.
Statistical models were estimated separately for first-born children and
children of higher birth order. This allowed authors to estimate the effects of
factors such as previous birth interval, death among previous children, and sex
combination of older siblings in a straightforward way. Table 6 shows the
descriptive statistics of the factors for the two groups of children by birth order.
Standard of living index was created by IIPS based on house type, toilet
facility, source of lighting, main fuel for cooking, source of drinking water,
whether the house has separate room for cooking, ownership of house, ownership
of agricultural land, ownership of irrigated land, ownership of livestock, and
ownership of 20 durable goods, classified as low, medium or high. Thirty per cent
of children of first-born children and 41 per cent of higher-order children were
born to households with low standard of living. The observed difference in the
standard of living by birth order is probably owing to the tendency for women in
poor families to have more children than women in better-off families.
Mothers’ level of education was classified in three categories: no formal
education, primary, or higher. Table 6 shows that mothers’ education is lower
among high-order births (63 per cent with no formal education) than among first
births (41 per cent with no formal education). This pattern is not surprising because
women with low level of education tend to have more children and thus, more
high-order births than the first births. In addition, mothers who gave first birth
during the five-year period before the survey are likely to be younger and have
higher level of education than those who gave birth to higher-order births.
30
Asia-Pacific Population Journal, Vol. 21, No. 1
Table 6. Descriptive statistics of the covariates, Indian children
born five years before 1998-1999 NFHS-2 survey
Covariates
Birth Birth order 2
All
order 1
and over
children
Household/mother-level factors
Standard of living index of household is low (per cent)
30
41
38
Mother had no formal education (per cent)
43
63
56
Mother's education is primary (per cent)
16
15
16
Mother's education is more than primary (per cent)
43
22
28
96.19
96.07
96.1
48
--
48
--
14
--
Child-level factors
Year of birth (mean)
Child is a girl (per cent)
Child is a girl and has no brothers (per cent)
Child is a girl and has at least one brother (per cent)
--
34
--
Mother's age at birth <18 (per cent)
28
--
--
Mother's age at birth <20 (per cent)
--
13
--
Preceding birth interval <24 months (per cent)
--
28
--
Birth order >4 (per cent)
--
25
--
Any death among older siblings (per cent)
--
10
--
Urban community
27
20
22
Percentage households with piped water (mean)
37
29
32
Percentage households with toilet facility (mean)
35
27
29
Percentage mothers with 2 tetanus vaccinations (mean)
72
65
67
Percentage children age 1 with full immunization (mean)
43
33
36
Community factors
Note: -- indicates not applicable
Indian women begin to have children at a very young age. Twenty-eight per
cent of first-born children were born to women below age 18. Among higher-order
births, 13 per cent were born to women below 20 years. Indian women also tend to
have short birth intervals. Twenty-eight per cent of children of birth order 2 or
higher were born less than 24 months after the birth of previous child. Substantial
proportion of births is of order 5 or higher reflecting high level of fertility. Ten per
cent of second or higher-order children are born to families with some experience
of child death.
Asia-Pacific Population Journal, April 2006
31
At the community level, the average prevalence of piped water and toilet
facility are about one third. The difference between first-born children and
higher-order births are probably owing to the fact that fertility tends to be higher in
less developed communities than in more developed communities.
Community-level prevalence of antenatal tetanus vaccinations averages about two
thirds at 67 per cent. By contrast, the average community-level prevalence of early
childhood vaccinations is only slightly over one third at 36 per cent.
Results
The hazard ratios (relative risk) estimated by the hazard model are shown in
table 7. Low standard of living increases early childhood mortality only for
children of birth order 2 or higher but not for first-born children. It is possible that
first-born children, being very precious to the family, receive special care from
parents and the standard of living has little effect on their survival during the first
two years of life after controlling for the effects of other factors such as mother’s
education, mother’s age at birth, and community factors. Mother’s education
above primary school level lowers early childhood mortality substantially but
primary school level education has no statistically significant effect on children’s
early childhood mortality. Relatively weak effect of primary level education of
mother on child mortality is commonly found in other studies as well (see for
example, Desai and Alva, 1998).
Many previous studies have documented high level of son preference and
consequent excess female child mortality in India (Arnold, Choe and Roy, 1998;
Das Gupta, 1987; Basu, 1989). Results from the present analysis are consistent
with these earlier studies: among the first-born children, early childhood mortality
is lower among girls as in most other populations, showing no evidence of
discrimination against daughters. Among higher-order births, girls experience
higher early childhood mortality, and the excess female mortality is more evident if
the girl has no surviving brothers, reflecting discrimination against daughters
especially when there are other daughters in the family.
Indian women begin their childbearing early as shown in table 6. The results
indicate that early childbearing is associated with increased early childhood
mortality. First-born children to mothers under age 18 experience 45 per cent
higher early childhood mortality risk than children born to older mothers. The
adverse effect of early childbearing continues with higher-order births as well,
although the relative risk is lower. Other factors associated with fertility behaviour
– previous birth interval and high birth order both have statistically significant
positive association with early childhood mortality as well. The relative risk of
early childhood mortality associated with short birth interval is especially high.
32
Asia-Pacific Population Journal, Vol. 21, No. 1
Children born to women who already experienced a death of children have higher
risks of early childhood mortality than those born to mothers who have not
experience any child death, consistent with findings from earlier studies.
Table 7. Relative risks of dying associated with covariates among
Indian children born five years before 1998-1999 NFHS-2 survey,
estimated from Cox model by birth order
Birth
order 1
Birth order 2
and over
Standard of living index of family is low
1.00
1.24*
Mother's education is primary
0.87
0.92
Mother's education is more than primary
0.62*
0.65*
Year of birth
0.98*
0.98*
Child is a girl
0.84*
Covariates
Household/mother level factors
Child-level factors
Child is a girl and has no brothers
Child is a girl and has at least one brother
Mother's age at birth <18
--
--
1.35*
--
1.16*
1.45*
--
Mother's age at birth <20
--
1.15*
Preceding birth interval <24 months
--
1.83*
Birth order >4
--
1.22*
Any death among older siblings
--
1.59*
Community factors
Urban community
1.09
1.07
Percentage households with piped water
0.78
1.02
Percentage households with toilet facility
0.67*
0.69*
Percentage mothers with 2 tetanus vaccinations
0.94
0.76*
Percentage children age 1 with full immunization
0.54*
0.67*
Notes: -- indicates not applicable
* indicates p<0.05
Turning now to the community-level factors, urbanity of the community and
proportion of children born in households with piped water have no effect on early
childhood mortality in India. The community-level prevalence of access to toilet
facility, however, has large statistically significant negative effect on early
childhood mortality. In the present models, community-level prevalence of
antenatal tetanus toxoid vaccine has statistically significant effect on early
Asia-Pacific Population Journal, April 2006
33
childhood mortality only for children of birth order two or higher. As discussed
earlier, this variable was used as a proxy for the utilization of preventive
reproductive health programmes. In India, utilization of antenatal care is often
related with complications associated with pregnancies, women experiencing such
complications being more likely to seek care. Thus, it may not be a good measure
for preventive reproductive health programmes. As expected, child immunization
coverage has strong negative effect on early childhood mortality.
Simulation
The impact of the factors on early childhood mortality can be seen clearly
when the predicted values of mortality are computed under different scenarios
specified by different hypothetical values. In table 8, cumulative predicted
probabilities of dying before the end of the 24th month are computed under a
selected set of scenarios. The scenarios consist of changing values of statistically
significant factors to the level observed in Kerala, where the under-five mortality is
lowest among the states of India.
For example, the following question can be raised: What would be the level
of early childhood mortality in all India if the proportion of women with more than
primary school education were equal to the level observed in Kerala? The mortality
from the estimated hazard model in table 7 can be estimated by changing the value
of “mother’s education is more than primary” to the value observed in Kerala,
leaving all other factors as observed. The results are shown in table 8, scenario (1).
The table shows that the cumulative mortality at the end of 24th month would be 64
among first-born children, a 20 per cent reduction from 80. It would be 65 among
higher-order births, a 22 per cent reduction from 83. Combining the first- and
higher-order births, as the weighted average using the observed distribution of
children by birth order, mortality for all children would be 65; that is a 21 per cent
reduction from 82. Similar exercises can be carried out by either changing the
value of one factor at a time, or a number of factors simultaneously. Separate
estimates can be calculated for children of birth order one and higher and for all
children using the weighted average of mortality estimated for first- and
higher-order births. If the scenario includes changes in fertility behaviour, the
weighted average is computed using the observed distribution of children by birth
order, or implied distribution as appropriate.
Table 8 shows that if the proportion of women with more than primary school
education in all India increased to the level of Kerala, the cumulative mortality at
the end of 24th month would be 21 per cent lower than the observed mortality
(scenario 1). According to the present data, 43 per cent of mothers of first-born
children in all India had more than primary school education compared to 92 per
34
Asia-Pacific Population Journal, Vol. 21, No. 1
cent in Kerala. It would be an enormous task to increase women’s education in all
India to match the level reached in Kerala. But if this could be achieved, the early
childhood mortality would fall substantially.
Table 8. Cummulative probability of dying by the end of 24th month
under selected scenarios, predicted by the estimated hazard models
Predicted cumulative Per cent reduction in
probability of dying by mortality relative to
end of 24th month
“no change” scenario
Scenario (a)
First
births
Other
All
First Other All
births births births births births
(0)
No change
80
83
82
NA
NA
NA
(1)
Percentage of women with more than
primary education
64
65
65
20
22
21
(2)
Percentage of children born in households
with low standard of living index
80
79
79
NA
5
4
(3)
Percentage of chidlren whose mothers age
at birth was very young (< 18 for birth
order 1, <20 for birth order 2 and over)
74
80
78
8
4
5
(4)
Percentage of children birth order >2 with
preceding birth interval <24 months
80
82
81
NA
5
1
(5)
Percentage of children birth order >4
80
76
77
NA
8
6
(6)
Community-level prevalence of
households with toilet facility
66
68
67
18
19
18
(7)
Community-level prevalnece of more than
2 antenatal tetanus vaccinations
80
79
79
NA
6
4
(8)
Community-level prevalence of complete
childhood immunization
67
72
71
17
14
14
(9)
All fertility factors
11
74
73
73
8
12
(10) All MCH factors
67
68
68
17
18
18
(11) All fertility and MCH factors
62
62
62
23
25
24
(12)
42
36
38
47
56
54
All statistically significant factors
except year of birth, sex of child, and
sex combination of older siblings
Notes: (a) Changes to match the situation in the state of Kerala, India.
NA: indicates not applicable
Reducing the proportion of households with low standard of living from the
level observed across India to the level of Kerala would result in some reduction of
early childhood mortality but only by 4 per cent (scenario 2). This is not surprising
because, as discussed earlier, the economic status of Kerala state is not drastically
different from that of India as a whole.
Asia-Pacific Population Journal, April 2006
35
Altering fertility behaviour of all Indian women to the pattern observed in
Kerala by reducing early childbearing, short birth intervals, and high order births
could reduce early childhood mortality by 5 per cent, 1 per cent and 6 per cent,
respectively (scenarios 3, 4 and 5). Altering all fertility behaviour simultaneously
would result in a 11 per cent reduction (scenario 9) in early childhood mortality,
according to this analysis.
Improving sanitary conditions of communities by increasing the proportion
of households with toilet facility to the level of Kerala could reduce early
childhood mortality by 18 per cent (scenario 6). Reducing early childhood
mortality through improving the sanitary conditions would take as long as
increasing mothers’ level of education.
Increasing utilization of reproductive and child health programmes by
increasing antenatal tetanus toxoid vaccines and child immunizations in all India to
the level observed in Kerala would reduce early childhood mortality by 4 per cent
and 14 per cent, respectively (scenarios 7 and 8). Changing both would result in 18
per cent reduction (scenario 10) in early childhood mortality.
Changing fertility behaviour and utilization of reproductive and child health
programmes simultaneously would reduce early childhood mortality by 24 per cent
(scenario 11). In addition, altering all of the factors that affect early childhood
mortality to the level observed in Kerala would result in a 54 per cent reduction in
early childhood mortality (scenario 12).
Discussion
The simulation exercises discussed in the previous section did not include
one major factor of early childhood mortality: sex of child and sex combination of
surviving older siblings of children of birth order two or higher. The sex of children
cannot be altered easily by population and health programmes. However, those
programmes can work towards eliminating son preference behaviour of parents in
taking care of their children through community-based communication and
education programmes. Eliminating mortality differentials by sex of children and
sex combination of older siblings would result in a 11 per cent reduction in early
childhood mortality.11
Although many conditions in Kerala are conducive to low level of under-five
mortality, there are some exceptions. One of them is the level of poverty as
discussed earlier. Another is the supply of piped water. In all India, average
percentage of households with piped water in a community is 40 per cent. In
Kerala, it is less than 20 per cent. It is worthwhile to note, however, that 61 per cent
of households in this state purify the water before drinking it, 77 per cent of them
36
Asia-Pacific Population Journal, Vol. 21, No. 1
using boiling as a purifying method (IIPS and ORC Macro, 2001). By contrast,
only 39 per cent of households in all India purify the water before drinking it, the
most common method of purifying being straining water by cloth (59 per cent).
Only 26 per cent of households in all India boil water for purification (IIPS and
ORC Macro, 2000). Those differences in the way water is treated before drinking
may explain why in the present analysis, the community-level prevalence of piped
water was not found to have a statistically significant effect on early childhood
mortality. Another possible explanation for the weak association is that an
important variable was omitted because data were not available (the variations in
personal hygiene behaviour such as washing hands before cooking/eating) may
interact with the prevalence of piped water.
In Kerala, although fertility level is below replacement, some aspects of
fertility behaviour are not conducive to low level of childhood mortality. One is the
prevalence of short birth interval. In all India, 28 per cent of children of birth order
two or higher were born in less than 24 months of the birth of the previous child. In
Kerala, the proportion is 21 per cent. This analysis found that short preceding birth
interval is a statistically significant factor associated with increased early
childhood mortality. Our simulation exercise on birth intervals (changing the
prevalence of short birth interval to the level of Kerala) does not show full potential
for reducing early childhood mortality through altering the child spacing
behaviour. How would early childhood mortality change if the short previous birth
intervals were eliminated? Table 9 shows the results of simulations from scenarios
including this hypothesis. If previous birth intervals of less than 24 months were
prevented (set the proportion to zero), the early childhood mortality would be
reduced by 6 per cent. If this scenario were combined with fertility behaviour
changes regarding early childbearing and prevalence of high order birth that
matched the levels of Kerala, the early childhood mortality would be reduced by 16
per cent. Those changes, combined with the improved coverage of antenatal
tetanus toxoid vaccines and child immunizations at the level observed in Kerala
would reduce early childhood mortality by 30 per cent.
According to the present statistical model, the standard of living index of
household has a weak association with early childhood mortality. Most studies,
especially those based on country-level analysis document that poverty is probably
the most important factor associated with high under-five mortality. Are the present
results contrary to this common pattern? The weak association between standard of
living index and early childhood mortality the authors found in India is likely to be
caused by the measurement of poverty that used the standard of living index. Poverty
leads to high level of mortality through malnutrition, poor access to medical care in
addition to the factors examined here. The standard of living index, based mostly on
Asia-Pacific Population Journal, April 2006
37
durable goods of the household is likely to be a poor measure of household
expenditure which would be more directly associated with those intermediate
variables. Furthermore, the standard of living index may have different meanings for
urban and rural households and in summary, may be a poor measure of poverty. When
a better measure of household-level poverty or expenditure is used, a stronger
association with the early childhood mortality may be found.
Table 9. Probability of dying at ages 0-24 under the assumption of
eliminating birth intervals < 24 months, predicted by
the estimated hazard models
Predicted
Per cent
cummulative reduction in
probability
mortality
of dying at relative to “no
ages 0-24
change”
months
scenario
Scenario (a)
(0) No change
82
NA
(1) Eliminate of children birth order >2 with preceding birth
interval <24 months
77
6
(2) Reduce children born to young mothers and children of
birth order >4 to the level observed in Kerala, and
eliminate children of birth order >1 with preceding birth
interval <24 months
69
16
(3) Reduce children born to young mothers and children of
birth order >4 to the level observed in Kerala, and
eliminate children of birth order >2 with preceding birth
interval <24 months, and increase antenatal tetanus toxoid
innoculation and children's immunization to the level
oberved in Kerala
58
30
Notes: (a) Changes to match the situation in the state of Kerala, India.
NA: indicates not applicable.
Goal 4 calls for a reduction of under-five mortality by two thirds (67 per cent)
by 2015. If the current trend in under-five mortality in India (13 per cent reduction
in six years) continues, one can expect a 54 per cent reduction by 2015; about 13
per cent short of the Goal. The present analysis shows that this gap can be closed by
effective family planning programmes resulting in reduction of early childbearing,
short birth intervals and high-order births, combined with increased utilization of
reproductive and child health programmes including antenatal tetanus vaccination
and child immunizations. Eliminating discrimination against girl children can
reduce early childhood mortality even further.
38
Asia-Pacific Population Journal, Vol. 21, No. 1
Whether findings based on the analysis of Indian data will apply in other
countries that need to improve child survival substantially remains to be validated
by additional evidences. In the meantime, it is encouraging to have an evidence that
early childhood mortality can be reduced substantially by family planning and
reproductive and child health programmes even under unfavourable conditions in
terms of poverty, women’s education and community-level sanitary conditions.
The country-level analysis showed that under-five mortality rate and
maternal mortality ratios are highly correlated and that they share common set of
determinants. Thus, the intervention programmes designed for reducing under-five
mortality rate are likely to reduce maternal mortality ratio as well.
Some countries may have inadequate technical and financial resources for
improving reproductive and maternal and child health services. International
cooperation both within the Asian and Pacific region and wider global community
may be required to meet the challenges of the Millennium Development Goals at
the regional level.
Acknowledgements
An earlier version of the paper was presented at the ad-hoc Expert Group
Meeting on the Implementation of the Plan of Action on Population and Poverty of
the Fifth Asian and Pacific Population Conference held at Bangkok from 8-10
November 2005. The authors are grateful to comments provided by Bhakta
Gubhaju and participants of the ad-hoc Expert Group Meeting. Computer
programming assistance provided by Gayle Yamashita, Computer Specialist at
East-West Center and research assistance provided by Katherine Miller, graduate
student in Public Health at the University of Hawaii are greatly appreciated.
Asia-Pacific Population Journal, April 2006
39
Endnotes
1. This article has been submitted to ESCAP as a theme study on Health and Millennium Development
Goals: Policies and Strategies to Meet the Millennium Development Goals of Reducing Child Mortality
and improving maternal health as part of the ESCAP-UNFPA project on Population, Poverty and
Development implemented in 2005.
2. The designation of regions follows the convention used by ESCAP, UNDP and ADB in this report A
Future Within Reach (2005). Thus, the analysis excludes countries in the Pacific region.
3. It should be noted here that because the indicator of progress towards the Goal is binary (0 or 1), the
correlation coefficients are not very large.
4. The constant rate of decline (equivalent to simple-interest approach) was used rather than the
constant relative rate of decline (equivalent to the compound-interest approach).
5. Some states show negative reduction in under-five mortality, indicating increase in the under-five
mortality. The reasons for the increase in mortality is not clear but the inaccuracy of data may explain
part of the unexpected result.
6. Equivalent to approximately US$ 376. As of January 1999, one U.S. dollar was equivalent to 42.6
Indian rupees.
7. Among women who gave births during the five-year period before the survey, for the last and
next-to-last births.
8. Among one-year olds at the time of survey.
9. Based on information of last birth and next-to-last birth for each woman.
10. Based on information from children born during the five-year period before the survey and
surviving at the time of survey.
11. The computation, not shown here, assumes that the early childhood mortality of female children
would be the same as that of male children, regardless of the sex combination of older siblings.
40
Asia-Pacific Population Journal, Vol. 21, No. 1
Appendix A1. Progress towards meeting Goals 4 and 5 as assessed in 2003 by ESCAP, UNDP
and ADB: countries in the ESCAP region with 2004 GNI ppp per capita < US$ 4,000
Goal 4
(under-5 and infant
mortality)
Country
Goal 5
(improve maternal
health)
Making
Already
Making
Already
slow
achieved or
slow
achieved or
expected to progress or expected to progress or
regressing
achieve
regressing
achieve
East and North-East Asia
Democratic People's Republic of Korea
X
X
Mongolia
X
X
South-East Asia
Cambodia
Indonesia
X
X
X
X
Lao People's Democratic Republic
X
Myanmar
X
X
Timor-Leste
X
X
Viet Nam
X
X
X
South and South-West Asia
Afghanistan
Bangladesh
X
X
X
X
Bhutan
X
India
X
X
Maldives
X
X
Nepal
X
X
Pakistan
X
X
Azerbajian
X
X
Georgia
X
X
Kyrgyzstan
X
X
Tajikstan
X
X
Uzbekistan
X
X
North and Central Asia
X
Sources: ESCAP, UNDP and ADB, 2005 (p. 13 for progress); PRB, 2005 for GNI ppp per capita.
Notes:
1. The table excludes the countries for which data on either Goal 4 progress or Goal 5
progress are not available.
2. The table includes Myanmar, Timor-Leste, Afghanistan, Bhutan, and Maldives for which
per capita income is unknown.
Asia-Pacific Population Journal, April 2006
41
Appendix A2. Progress towards meeting Goals 4 and 5 as assessed in 2003 by ESCAP, UNDP
and ADB: countries and areas in the ESCAP region
with 2004 GNI ppp per capita >= US$ 4,000
Country/area
Goal 4
Goal 5
(under-5 and infant
mortality)
(improve maternal
health)
Making
Already
Making
Already
slow
achieved or
slow
achieved or
expected to progress or expected to progress or
regressing
achieve
regressing
achieve
East and North-East Asia
China
X
Hong Kong, China
X
X
X
Macao, China
X
X
Japan
X
X
Republic of Korea
X
X
South-East Asia
Brunei Darussalam
X
X
Malaysia
X
X
Philippines
X
X
Singapore
X
X
Thailand
X
X
South and South-West Asia
Iran (Islamic Republic of )
X
X
Sri Lanka
X
X
Turkey
X
X
X
X
North and Central Asia
Armenia
Kazakhstan
Russian Federation
Turkmenistan
X
X
X
X
X
X
Sources: ESCAP, UNDP and ADB, 2005 (p. 13 for progress); PRB, 2005 for GNI ppp per capita.
Notes: 1. The table excludes the countries for which data on either Goal 4 progress or Goal 5
progress are not available.
2. The table includes Brunei Darussalam for which per capita income is not available.
42
Asia-Pacific Population Journal, Vol. 21, No. 1
References
Academy of Preventive Medicine, Kazakhstan and Macro International Inc. (1999). Kazakhstan
Demographic and Health Survey 1999. Claverton, Maryland, USA: Academy of Preventive
Medicine and Macro International Inc.
Ahmed, O. B., A. D. Lopez, and M. Inoue (2000). “The decline in child mortality: A reappraisal”,
Bulletin of the World Health Organization, No. 78, pp. 1175-1191.
Arnold, Fred, Minja K. Choe and T.K. Roy (1998). “Son preference, the family-building process and
child mortality in India”, Population Studies, No. 52, pp. 301-315.
Basu, Alaka M. (1989). “Is Discrimination in Food Really Necessary for Explaining sex Differentials in
Childhood Mortality?”, Population Studies, No. 43, pp. 193-210.
Black, Robert E., Saul S. Morris, and Jennifer Bryce (2003). “Where and why are 10 million children
dying every year?”, The Lancet, No. 361, pp. 2226-2234.
Committee for Population, Family and Children, Viet Nam and ORC Macro (2003). Vietnam
Demographic and Health Survey 2002. Calverton, Maryland, USA: Committee for
Population, Family and Children and ORC Macro.
Das Gupta, Monica (1987). “Selective Discrimination Against Female Children in Rural Punjab,
India”, Population and Development Review, vol. 13, pp.77-100.
Desai, Sonalde and Soumya Alva (1998). “Maternal Education and Child Health”, Demography, vol.
35, No. 1, pp. 71-81.
Economic and Social Commission for Asia and the Pacific (ESCAP), United Nations Development
Programme (UNDP) and Asian Development Bank (ADB) (2005). A Future Within Reach:
Reshaping Institutions in a Region of Disparities to Meet the Millennium Development
Goals in Asia and the Pacific. New York: United Nations.
International Institute for Population Sciences (IIPS) (1995). National Family Health Survey (MCH
and Family Planning), India 1992-93. Bombay: IIPS.
International Institute for Population Sciences (IIPS) and ORC Macro (2000). National Family Health
Survey (NFHS-2), 1998-99: India. Mumbai: IIPS.
International Institute for Population Sciences (IIPS) and ORC Macro (2001). National Family Health
Survey (NFHS-2), India, 1998-99: Kerala. Mumbai: IIPS.
International Monetary Fund (IMF), Organization for Economic Co-operation and Development
(OECD), United Nations (UN) and World Bank Group (WBG). 2000. 2000 A Better World
for All: Progress towards the International Development Goals.
Koenig, Michael A., James F. Phillips, Oona M. Campbell and Stan D’Souza (1990). “Birth intervals
and childhood mortality in rural Bangladesh”, Demography, vol. 29, No. 2, pp. 251-265.
Miller, Jane E. (1991). “Birth intervals and perinatal health: An investigation of three hypotheses”,
Family Planning Perspectives, No. 23, pp. 62-70.
Asia-Pacific Population Journal, April 2006
43
Miller, J.E., James Trussell, Anne R. Pebley and B. Vaughan (1992). “Birth spacing and child mortality
in Bangladesh and Philippines”, Demography, vol. 29, No. 2, pp. 305-318.
Ministry of Health and Family Welfare, Government of India (1992). Action Plan for Revamping the
Family Welfare Programme in India. New Delhi, India: Ministry of Health and Family
Welfare.
Ministry
of Statistics and Programme Implementation, Government of India (2005).
http://mospi.nic.in/10_percapnsdp_cor_9394ser.htm accessed on 11 October 2005.
Mosley, W. Henry and Lincoln C. Chen (1984). “An analytical framework for the study of child
survival in developing countries”, Population and Development Review, 10 Suppl, pp.:
25-45.
Office of Registrar General, India (2005). http://www.censusindia.net/t_00_003.html accessed 21
November 2005.
Pandey, Arvind, and others (1998). Infant and Child Mortality in India. National Family Health Survey
Subject Reports, Number 11. Mumbai and Honolulu: IIPS and EWC POP.
Population Reference Bureau (2005). Women of Our World. Washington, D.C., U.S.A.: Population
Reference Bureau.
__________ (2005). World Population Data Sheet. Washington, D.C., U.S.A.: Population Reference
Bureau.
Research Institute of Obstetrics and Pediatrics, Kyrgyz Republic and Macro International Inc. (1998).
Kyrgyz Republic Demographic and Health Survey, 1997. Claverton, Maryland, USA:
Research Institute of Obstetrics and Pediatrics, Ministry of Health of the Kyrgyz Republic
and Macro International Inc.
Setty-Venugopal, V. and U. D. Upadhyay (2002). “Birth spacing: three to five saves lives”, Population
Reports, Series L, No. 13. Baltimore: Johns Hopkins Bloomberg School of Public Health,
Population Information Program.
Tulloch, J. (1999). “Integrated Approach to Child Health in Developing Countries”, Lancet, No. 354
(Suppl. 1), pp. 7-24.
Tsui, A. O., J. N. Wasserheit, and J. G. Haaga, eds. (1997). “Healthy Pregnancy and Childbearing,” in
Reproductive Health in Developing Countries: Expanding Dimensions, Building Solutions.
Washington, D.C.: National Academy Press.
UNICEF, UNFPA and WHO (2004). Maternal Mortality in 2000: Estimates Developed by SHO,
UNICEF, and UNFPA. Geneva: World Health Organization.
Winikoff, Beverly (1983). “The effects of birth spacing on child and maternal health”, Studies in
Family Planning, vol. 14, No. 1, pp. 231-245.
World
44
Health
Organization
(2002).
http://www.who.int/whr2002.htm.
Reducing
Risk,
Promoting
Healthy
Life.
Asia-Pacific Population Journal, Vol. 21, No. 1
Readiness, Willingness and
Ability to Use Contraception
in Bangladesh
This study has established that most women now want to control
their fertility and consider fertility control as normatively acceptable,
as well as convenient in terms of availability/accessibility and cost.
By Haider Rashid Mannan and Roderic Beaujot*
In his frequently quoted article, Coale (1973) proposes that one weakness of
the demographic transition theory is that it indicates a high degree of
modernization as sufficient to cause a fall of fertility, without indicating the degree
of modernization that is necessary. By summarizing the findings of historical
* Haider Rashid Mannan, Doctoral candidate, School of Population Health, University of Western
Australia, Nedlands, WA 6009, Australia, e-mail: [email protected] and Assistant Professor,
Department of Statistics, University of Dhaka, Dhaka, Bangladesh and Roderic Beaujot, Professor,
Department of Sociology, University of Western Ontario, London, Ontario N6A 5C2, Canada, e-mail:
[email protected].
Asia-Pacific Population Journal, April 2006
45
studies of European communities, Coale proposed three broad conditions
necessary for fertility transition. He argued that modernization ultimately
establishes these conditions, but that they can also occur in communities that have
undergone little modernization. Lesthaeghe and Vanderhoeft (1998) later
described the three conditions for fertility transition under the heading “readiness”,
“willingness” and “ability”. Economic readiness means that fertility control must
be advantageous to the actor so that fertility is within the calculus of conscious
choice. Willingness means that fertility control must be legitimate and normatively
acceptable. The basic question addressed by “willingness” is to what extent
fertility control runs counter to established traditional beliefs and codes of conduct,
and to what extent there is a willingness to overcome objections and fears. Ability
refers to the availability and accessibility of contraceptive techniques. Similarly,
Ahmed (1987: 363), applying Easterlin’s supply-demand theory of fertility
observes that “studies on contraceptive use most often view three
variables-motivation, attitude, and access-as the key determinants”. Motivation
stems from having or expecting to have too many children or having them too soon.
Although this has similarity with Coale’s notion of “economic readiness”, it does
not necessarily capture whether fertility control is economically advantageous to
an individual. In this paper the authors name Easterlin’s notion of motivation as
simply “readiness” to distinguish it from Coale’s broader notion of “economic
readiness”. Attitude refers to broad notions of acceptability of family planning in
general and feelings about specific contraceptive methods in particular and is
similar to Coale’s notion of willingness. Access or the “costs of fertility
regulation”, as described by Easterlin (1975) pertains to the availability of
contraceptives and selected services and is similar to Coale’s notion of ability.
Coale’s “economic readiness” and “willingness” conditions relate
respectively to economic and cultural dimensions of an innovation. Thus, we
expect a greater economic readiness to use contraception for individuals who
have more living children, and who have more contact with modern economic
conditions where it is advantageous to have fewer children. By contrast,
persons who are in more traditional environments are expected to have less
willingness to use contraception. Some authors have argued in particular that
normative changes have been based on the diffusion of new forms of family
behaviour (Knodel, 1974; Lesthaeghe, 1977; Cleland, 1985; Rosero-Bixby
and Casterline, 1993). Those changes fall under the category of willingness in
terms of Coale’s notion. Knowledge about family planning methods plays an
important role in the diffusion process. Coale’s precondition of ability to
regulate fertility includes family planning knowledge and helps to explain the
diffusion process well (Knodel, 1974).
46
Asia-Pacific Population Journal, Vol. 21, No. 1
In general, fertility has declined much more rapidly in Asia and Latin
America than in European countries during the demographic transition. In
Bangladesh, for example, fertility has declined significantly over a relatively short
period of time through massive adoption of modern contraceptives in spite of
minimal economic development (Stevens, 1994). There has been debate about
whether the Bangladesh fertility decline has been achieved solely owing to
massive adoption of modern contraceptives (Carty, Yinger and Rosov, 1993;
Cleland and others, 1994; Caldwell and others, 1999). The present study does not
attempt to resolve this debate. Instead, it will seek to determine the extent to which
questions of readiness and willingness to control fertility are prevailing in the
Bangladeshi society following the fertility decline, besides the question of
availability and accessibility to contraception. Although ideally it would have
been important to determine the extent to which questions of economic readiness
are prevailing in the Bangladeshi society following the fertility decline, because
of difficulty in proper operationalization of this concept, the study will examine a
similar concept; motivation or simply readiness to regulate fertility as described
by Easterlin (1975).
Traditional analyses of determinants of contraceptive use have often
examined socio-economic and demographic factors as potential determinants
of contraceptive use. Those variables have been found to be important
determinants of contraceptive use in Bangladesh (Ullah and Chakraborty,
1993; Khan and Rahman, 1997). However, very few studies have examined the
effects of readiness, willingness and ability on contraceptive use. Previous
researches have examined the relevance of some of the aspects of
“willingness” to contraceptive method choice, by measuring the perception of
side effects and inconvenience of use (Mannan, 2002). But, the relevance of
those aspects and other aspects of “willingness” have not been examined with
regard to contraceptive use. Only a single study by Ahmed (1987) examined
the impact of motivation or readiness to regulate fertility on contraceptive use,
in which it was found that strong motivation to regulate fertility led to higher
contraceptive use in rural Bangladesh. Previous researches in Bangladesh
have shown that contraceptive access, quality of field workers and quality of
clinical services encourage a greater adoption and continuation of
contraception (Koenig, 2003; Islam, Barua and Bairagi, 2003) while
community contraceptive availability encourages greater adoption (Tsui and
others, 1981). Thus, the present study seeks to determine the relative
importance of readiness, willingness, ability and socio-economic,
sociocultural and demographic variables with regard to contraceptive use in
Bangladesh.
Asia-Pacific Population Journal, April 2006
47
Data, measures and methods
The authors used data from the 1996-1997 Bangladesh Demographic and
Health Survey (BDHS), which is a national survey conducted under the authority
of the National Institute for Population Research and Training (NIPORT) of the
Ministry of Health and Family Welfare (Mitra and others, 1997). Using a two stage
stratified sample design a total of 9,355 ever-married women aged 10-49 were
identified as eligible for the individual interview. Of those, 9,127 or 97.8 per cent
were successfully interviewed. The sample is first restricted to the 8,450 currently
married women aged 10-49 for whom information was obtained on contraceptive
usage. In order to consider solely women for whom current use of contraception is
relevant, the analysis is based on the 7,517 currently married fecund women who
were not pregnant at the time of the survey. Current use of contraception was used
as the dependent variable for the study of contraceptive use because it allows for
the analysis of behaviour at the time of the interview and it is less subject to error.
Future intention to use contraception would also have been an important dependent
variable for the purpose of the study. Lack of data did not allow for the inclusion of
this variable in the study.
The ready, willing and able variables were measured as dummy variables
following the guidelines by Lesthaeghe and Vanderhoeft (1998). Readiness was
measured by concentrating exclusively on subsequent births. In the BDHS
questionnaire, non-pregnant women were asked whether they preferred to have
a/another child or preferred not to have any more children (Mitra and others, 1997:
221). In addition, women who wanted a/another child were asked when they would
like to have the next child. Those who intended to wait for two or more years before
having a/another child, or who did not want more children, were considered as
being ready for using contraception. Women who wanted a child within two years,
or were undecided about their future fertility were considered as being not ready
for using contraception. The measure for readiness was based on 7,497 cases.
Based on this measure, 82.7 per cent were ready and 17.3 per cent were not ready to
use contraception.
The notion of “willingness” refers to considerations of legitimacy and
normative acceptability of family planning methods as well as willingness to
overcome objections and fears associated with family planning methods. In the
BDHS, women were asked whether they approved or disapproved of family
planning (Mitra and others, 1997: 222). In addition, they were asked to specify the
main reason for not intending to use a method. Women who disapproved of family
planning and did not intend to use contraception because of personal opposition,
opposition from husband, opposition from others, religious prohibitions, fears for
48
Asia-Pacific Population Journal, Vol. 21, No. 1
health which included answers like bad for health, side effects, inconvenient to use,
and interference with body’s normal processes, were regrouped in the category
“non-willingness”. The measure for willingness was based on 7,517 cases, of
which 95.3 per cent were willing to use contraception.
The notion of “ability” refers to the accessibility of family planning methods.
In the BDHS, women were asked which methods they had heard about (Mitra and
others, 1997: 205, 211). They were also asked whether they knew of any source
from where they could obtain a family planning method. Women who did not
intend to use contraception were asked the main reasons for not doing so, the
reasons included among others, difficulty of access to family planning services and
difficulty pertaining to costs. Respondents who had no knowledge about methods
of contraception, no knowledge about family planning services, had difficulty in
accessing family planning services or for whom family planning services were too
distant, and had difficulty pertaining to costs were classified as “non-able”, which
amounted to 6.3 per cent of the measure for ability which was based on 7,517 cases.
The questionnaire simply asked for the main reason, rather than multiple reasons,
for not intending to use contraception. This may result in the overestimation of
proportions willing and able. In addition, it would have been better to measure
willingness and ability as continuous variables by counting the number of “yes”
responses to the reasons related to willingness and ability, but respondents were
only asked to specify through a single question the main reason for not intending to
use contraception and thus it was not possible to measure those variables in this
manner. Also, given the politically sensitive questions, it is difficult to know
whether the interview setting may have had any influence on answers to the
questions which were related to the measures of willingness and ability.
For testing the association between each covariate and the outcome at the
bivariate level, the chi-square test was used. Logistic regression is used for
multivariate analysis, with the categorical and interval variables indicated in the
tables (see tables 1 and 2). The interval variables were also entered as squared
terms to capture their non-linear effects on the dependent variable. To avoid
problems of multicollinearity, the linear and squared terms for the interval
variables were centred. Since the authors’ main objective is to examine the
individual effects of readiness, willingness and ability on contraceptive use, their
interactions were not considered for easier interpretability of results. The model
chi-square statistic has been used for evaluating model fit. It is a likelihood ratio
(LR) test between -21ogLR when only the constant term is in the model and
-21ogLR when the constant and the covariates are in the model. A LR test can also
be used to compare the fit of two nested models. The statistical significance for
Asia-Pacific Population Journal, April 2006
49
each covariate is tested using Wald’s statistic. Each covariate is tested at 5 per cent
level of significance. For obtaining national estimates based on BDHS, appropriate
weights were derived to adjust for oversampling from urban population and
non-response (Mitra and others, 1997). All analyses in the present study will thus
consider weights.
Sample characteristics and descriptive results
The distributions of the socio-economic, sociocultural, residence/region
and demographic variables are shown in tables 1 and 2. The only sociocultural
variable which was included in this study is religion. Lack of data did not allow
for the consideration of other relevant sociocultural factors such as
superstitions or taboos against family planning (eg., folk stories), religiosity,
etc. The proportions currently using contraception, ready, willing and able
within categories of the variables are also shown in tables 1 and 2. Regarding
ready, willing and able, there is no large variation between urban and rural
areas, but the Sylhet region shows markedly lower values than the other
divisions. While the differences are not large, it appears that religion affects
willingness more than the two other variables since religion has statistically
significant effect only on willingness. There is markedly lower readiness to
use contraception for women who are young, recently married or have no
children, but those demographic considerations have much less impact on
willingness and ability to use contraception (table 2). Similarly, the lack of
living sons seems to affect readiness much more than willingness. Women
with secondary or more education show low readiness but high willingness and
ability to use contraception.
Contraceptive use increases curvilinearly with current age, number of living
children, number of living sons, and duration of marriage. It increases with the
educational levels of women and their husbands. Urban women have higher
contraceptive use rates than their rural counterparts while women of Sylhet
division have considerably lower rates than women of other divisions.
Predictors of readiness, willingness and ability to use contraception
While the proportions who do not qualify as willing and able to use
contraception are on the low side for analysis through logistic regression, it is
still worth observing the extent to which those ready, willing and able
variables can be predicted by the socio-economic, sociocultural and
demographic variables (table 3). The results are discussed below.
50
Asia-Pacific Population Journal, Vol. 21, No. 1
Table 1. Currently married, fecund and non-pregnant women, showing percentage who are
currently using contraception and are ready, willing and able to use contraception, by RWA,
region/residential, sociocultural and socio-economic variables, Bangladesh, 1996-1997
Characteristics
Number of
Using
cases
contraceptive
Ready
Willing
Able
Readiness
Not ready
Ready
1,294 (17.3)
6,203 (82.7)
15.1
63.7
p<0.0001
---
---
---
Willingness
Not willing
Willing
352 94.7)
7,164 (95.3)
8.5
57.6
p<0.0001
---
---
---
Ability
Not able
Able
474 (6.3)
7,043 (93.7)
8.6
58.5
p<0.0001
---
---
---
Place of residence
Urban
Rural
873 (11.6)
6,644 (88.4)
63.7
53.6
p<0.0001
84.4
82.5
p>0.05
97.9
95.0
p<0.001
95.9
93.4
p<0.01
Region of residence
Sylhet
Barisal
Chittagong
Dhaka
Khulna
Rajshahi
373 (5.0)
492 (6.5)
1,489 (19.8)
2,370 (31.5)
929 (12.4)
1,963 (24.8)
24.7
56.1
42.5
55.8
68.1
64.5
p<0.0001
68.0
84.9
80.9
83.2
85.0
84.9
p<0.0001
82.3
96.3
93.6
96.0
96.4
97.6
p<0.0001
80.5
92.7
92.9
93.3
96.2
96.5
p<0.0001
Women's education
No schooling
Primary incomplete
Primary complete and above
3,958 (52.7)
2,102 (28.0)
1,456 (19.4)
52.1
56.8
62.0
p<0.0001
83.8
82.7
79.9
p<0.01
93.5
96.2
98.9
p<0.0001
91.8
94.4
97.7
p<0.0001
Current work status
Not working
Working
5,106 (67.9)
2,408 (32.0)
52.5
61.3
p<0.0001
81.4
85.7
p<0.0001
94.9
96.1
p<0.05
92.2
96.8
p<0.0001
Religion
Muslim
Non-Muslim
6,715 (89.3)
801 (10.7)
54.2
64.6
p<0.0001
82.6
83.9
p>0.05
95.1
97.0
p<0.05
93.6
94.8
p>0.05
.../
Asia-Pacific Population Journal, April 2006
51
Table 1. (Continued)
Number of
Using
cases
contraceptive
Characteristics
Ready
Willing
Able
Husband's education
No schooling
Primary incomplete
Primary complete
Secondary and above
3,269 (43.5)
1,872 (24.9)
1,590 (21.2)
722 (9.6)
51.9
53.3
58.7
69.4
p<0.0001
84.6
83.0
83.2
81.5
p>0.05
93.7
95.5
96.1
98.8
p<0.0001
91.8
93.5
95.7
98.9
p<0.0001
Husband's occupation
Professional/business
Others
1,933 (25.7)
5,649 (74.3)
63.6
55.2
p<0.0001
84.0
82.2
p>0.05
96.7
94.8
p<0.001
95.8
93.0
p<0.0001
55.3
82.7
95.3
93.7
Total
7,517
Note: The percentages on the parentheses may not add up to 100.00 because of rounding. The
percentages for ready are based on 7,497 cases. The p-values are based on chi-square test.
Table 2. Currently married, fecund and non-pregnant women, showing percentage who are
currently using contraception and are ready, willing and able to use contraception, by
demographic variables, Bangladesh, 1996-1997
Characteristics
Number of
cases
Using
contraceptive
Ready
Willing
Able
18.2
38.5
48.0
57.8
66.9
68.9
62.9
47.6
44.2
68.2
77.0
81.7
89.0
92.0
95.2
98.4
95.0
96.2
96.6
96.5
96.1
93.9
93.3
87.6
76.0
90.6
94.6
95.9
96.6
94.8
91.4
86.3
p<0.0001
p<0.0001
p<0.0001
p<0.0001
Current age
Below 15
15-19
20-24
25-29
30-34
35-39
40-44
45-49
121 (1.6)
1,067 (14.2)
1,484 (19.7)
1,646 (21.9)
1,222 (16.3)
955 (12.7)
642 (8.5)
380 (5.1)
Number of living children
0
770 (10.2)
21.4
234.3
94.3
84.0
1
1,663 (19.5)
47.1
72.9
96.5
94.5
2
1,642 (21.8)
63.8
86.7
96.6
96.0
3
1,316 (17.5)
64.3
92.1
96.3
96.0
4
1,001 (13.3)
64.6
96.3
94.8
94.8
5 or more
1,326 (17.7)
57.6
97.2
92.4
92.4
p<0.0001
p<0.0001
p<0.0001
p<0.0001
.../
52
Asia-Pacific Population Journal, Vol. 21, No. 1
Table 2. (Continued)
Characteristics
Number of
cases
Using
contraceptive
Ready
Willing
Able
Number of living sons
0
2001 (26.6)
38.2
58.5
95.2
90.4
1
2552 (33.9)
60.1
86.8
96.7
95.8
2
1654 (22.0)
65.3
94.7
95.3
95.0
3
801 (10.7)
62.3
95.8
93.3
93.4
4
355 (4.7)
57.0
98.6
92.4
93.8
5 or more
153 (2.0)
49.7
96.1
90.3
88.9
p<0.0001
p<0.0001
p<0.0001
p<0.0001
Child loss experience
0
5042 (67.1)
55.5
80.1
96.6
94.1
1
1550 (20.6)
58.0
87.5
94.2
94.1
2
598 (7.9)
52.0
89.3
92.0
91.8
3 or more
328 (4.4)
46.0
89.0
86.3
88.1
p<0.0001
p<0.0001
p<0.0001
p<0.0001
Duration of marriage (years)
Below 5
1257 (16.9)
36.1
63.2
96.3
89.3
5-9
1368 (18.2)
48.6
76.7
96.5
95.2
10-14
1474 (19.6)
57.9
83.2
96.1
96.1
15-19
1182 (15.7)
66.7
89.0
96.2
96.4
20-24
929 (12.4)
67.7
92.2
95.3
95.0
25 or more
Total
1227 (16.3)
7517
59.4
95.6
91.1
90.2
p<0.0001
p<0.0001
p<0.0001
p<0.0001
55.3
82.7
95.3
93.7
Note: The percentages on the parentheses may not add up to 100 because of rounding. The percentages
for ready are based on 7,497 cases. The p-values are based on chi-square test.
With regard to readiness, the significant predictor variables are region of
residence, women’s education, husband’s education, number of living children,
number of living sons and marital duration. Women are significantly more ready
to use contraception with increase in their and their husband’s educational levels.
This contradicts the results of bivariate analysis. It was observed by bivariate
analysis that the relationship between readiness and education (for both women’s
and their spouse’s) changes with number of living children (results not given).
Women with fewer living children are less ready while at the same time they are
more educated. Among the regions, women of Chittagong division are least ready
to use contraception followed by women of Sylhet division. As was expected,
individuals with more living children are significantly more ready to use
Asia-Pacific Population Journal, April 2006
53
contraception. The results also show that individuals with more living sons are
significantly more ready to use contraception. This result was expected, given the
importance of son preference in situations where women are economically and
socially dependent on men.
As expected, religion and region of residence are also strong predictors of
willingness. Women of Sylhet division are least willing to use contraception
followed by women of Chittagong division. The other predictors of willingness are
women’s education and number of dead children.
For able the significant predictors are region of residence, women’s
education, husband’s education, women’s current work status, maternal age and
number of living children. Ability increases significantly with education. This is
expected because educated people are likely to have more knowledge about
contraceptive methods. Working women are significantly more able to use
contraception than their non-working counterparts. Those women are likely to
have more knowledge about contraceptive methods possibly through greater social
interaction. Women of Sylhet division are least able to use contraception. This
means that this region is lagging behind other regions of the country in terms of
availability/accessibility and knowledge of modern contraceptive methods. Other
researches have also shown that family planning programmes have been
implemented differently in the various regions (NIPORT, 2001; Freedman, Khoo
and Supraptilah, 1981), which affects the notion of ability.
It is noteworthy that religion is a strong predictor of willingness to use
contraception, with less willingness among Muslim women, but religion is a weak
predictor of readiness and ability to use contraception. By contrast, readiness is
much more affected by the number of living children, living sons, duration of
marriage and some socioeconomic variables. Thus, the results confirm the greater
importance of cultural factors to willingness, and socio-economic and
demographic factors to readiness to use contraception.
Table 3. Logistic regression for readiness, willingness and ability to use
contraception, Bangladesh, 1996-1997
Variables
Residence/regional variables
Place of residence (Urban)
Rural
Region of residence (Sylhet)
Barisal
Ready
Coefficient
Willing
Coefficient
Able
Coefficient
--0.2337
-1.2190*
--0.472
-1.5460*
--0.0058
-0.9978*
.../
54
Asia-Pacific Population Journal, Vol. 21, No. 1
Table 3. (Continued)
Variables
Ready
Coefficient
Willing
Coefficient
Able
Coefficient
Chittagong
Dhaka
Khulna
Rajshahi
0.5501**
1.1761*
1.3743*
1.4885*
1.0345*
1.6680*
1.6810*
2.2516*
1.1349*
1.2928*
1.9004*
1.9938*
Sociocultural variables
Religion (Muslim)
Non-Muslim
-0.0977
-0.5672***
-0.0821
Socio-economic variables
Women's education (No education)
Primary incomplete
Primary complete +
Current work status (Not working)
Working
Husband's occupation (Prof./business)
Others
Husband's education (No education)
Primary incomplete
Primary complete
Secondary+
-0.1142
0.2569***
-0.11
-0.0517
-0.1353
0.2222
0.3570***
-0.3301***
1.3870*
-0.2456
--0.1435
-0.2361
0.1688
0.2971
-0.4455*
1.0993*
-0.7837*
--0.1705
-0.1324
0.273
1.0937*
Demographic variables
Current age
Current age squared
Number of living children
Number of living children squared
Duration of marriage
Duraction of marriage squared
Number of living sons
Number of living sons squared
Child loss experience
Child loss experience squared
Constant
-0.086
0.0016
1.7272*
-0.1507*
-0.2112*
0.0057*
0.7093*
-0.0979**
-0.0592
-0.0033
-0.1777
0.1552
-0.0026
0.2192
-0.0173
-0.0564
0.0007
0.1776
-0.0345
-0.2253***
0.0057
-0.2303
0.2885*
-0.0043*
0.4112*
-0.0343*
-0.0302
-0.0004
0.1434
-0.0275
-0.1422
0.0097
-3.5212*
303.239
24
7,496
449.676
24
7,496
Model Chi-square
Degrees of freedom
Sample size
1,859.981
24
7,496
Note: The reference category for a categorical variable is in the parentheses.
* indicates p<0.001; ** indicates p< 0.01; and *** indicates p<0.05.
Asia-Pacific Population Journal, April 2006
55
Relative importance of “ready, willing and able” compared to
other predictors of contraceptive usage
The bivariate analysis indicated that contraceptive use increases
curvilinearly with current age, number of living children, number of living sons,
and duration of marriage. Thus, squared terms are introduced in the logistic
models to capture their non-linear relationships with contraceptive use. Three
models were used for predicting contraceptive usage: Model A includes the terms
for ready, willing and able, Model B has all the socio-economic, sociocultural and
demographic variables but no ready, willing and able variables, and Model C is
the full model which has all the variables included in Models A and B (see table 4).
Based on LR tests all the models are found to be statistically significant indicating
a good fit for each model. Since the sample sizes for the three models are similar
and Model A is a subset of Model C, a LR test can be used to compare their fit.
Similarly, Model B and Model C can be compared by a LR test. A comparison
with Model A indicates that when the socio-economic, sociocultural and
demographic variables are introduced in Model C, the model chi-square increases
by 654.516 with 24 degrees of freedom (p=0.000). A comparison between Models
B and C indicates that model chi-square increases by 1035.935 with 3 degrees of
freedom (p=0.000) for the inclusion of ready, willing and able variables to the
model with only socio-economic, sociocultural and demographic variables as
covariates (Model B). Thus, the variables ready, willing and able together are
considerably more important than the socio-economic, sociocultural and
demographic variables in terms of model fit. A comparison between models with
and without controls for the socio-economic, sociocultural and demographic
variables further indicates that the coefficients for ready, willing and able are
fairly robust as they only differ in the first or second decimal points. On the other
hand, when the ready, willing and able variables are controlled in Model C, the
magnitude of the effects of most socio-economic, sociocultural and demographic
variables decreases considerably. This suggests that much of the effects of the
socio-economic, sociocultural and demographic variables on contraceptive use
are absorbed by the ready, willing and able variables. Thus, the effects of the
socio-economic, sociocultural and demographic variables on contraceptive use
are influenced to a large degree by the three variables. While these may seem to
suggest that the ready, willing and able variables act as intervening variables
between the more distant background variables and the outcome of contraceptive
usage, changes in those background characteristics are key to changes in the
readiness, willingness and ability to use contraception as has been found in the
previous section.
56
Asia-Pacific Population Journal, Vol. 21, No. 1
The results indicated that women who are ready to control fertility are
associated with increasing likelihood to use contraceptive than those who are not
ready to regulate fertility. This is in accordance with Ahmed’s (1987) findings who
had found, using Easterlin’s framework, that strong motivation to regulate fertility
led to higher contraceptive use in rural Bangladesh. It was found that women who
are able or have the means to control fertility are associated with increasing
likelihood to use contraceptive than those who are not able to regulate fertility.
Ahmed (1987) had similarly found earlier that lower costs of fertility regulation led
to higher contraceptive use in rural Bangladesh. However, he considered just one
indicator; distance to family planning clinics as a proxy for cost of fertility
regulation or means to regulate fertility while the present study considered all
issues relevant to means to regulate fertility. The study found that women who are
willing to regulate fertility are associated with higher likelihood to use
contraceptive than those who are not willing to regulate fertility. None of the
previous studies in Bangladesh including the one by Ahmed (1987) examined the
impact of normative willingness on contraceptive use. A few studies have
examined the influence of individual’s religious beliefs and religiosity on
contraceptive use (Bernhardt and Uddin, 1990; Kamal and Slogget, 1993), in
which it was found that those variables do not pose a significant barrier to
contraceptive use. However, individual’s religious beliefs and religiosity are only a
few predictors of willingness to regulate fertility (the other predictors are likely to
be superstitions against family planning, social conservatism, etc.) and thus cannot
fully explain individual’s willingness to regulate fertility.
The authors further found that urban, better educated, non-Muslim, and
currently working women are associated with higher odds to practice contraceptive
than those who are rural, less educated, Muslim, and not currently working. Wives
of professionals/businessmen are more likely to practice contraception than those
whose husbands are engaged in other occupations. In Bangladesh, women who
belong to the former category are likely to have higher socio-economic status.
There is also a strong regional variation in current use. Women of Sylhet and
Chittagong divisions have considerably lower use rates than those of other
divisions. It has been suggested that women of those two divisions are religiously
more conservative and have traditional values regarding family formation (Khan
and Raeside, 1998). The present study also supported this as women of those two
divisions were found to be less willing to use contraception than those of other
divisions. Low levels of education and low former-sector employment may also
partly contribute to the low use of contraceptives among Sylhet women. Both
bivariate and multivariate analyses in this study suggested that women of Sylhet
division are significantly less ready and able to regulate fertility than those of other
Asia-Pacific Population Journal, April 2006
57
divisions. It has been found that family planning programmes have been
implemented differently across the various regions of the country and Sylhet and
Chittagong divisions are lagging behind other divisions in terms of family planning
service delivery (NIPORT, 2001; Freedman, Khoo and Supraptilah, 1981).
Regardless of whether the ready, willing and able variables are controlled in
the analysis, the likelihood of contraceptive use is higher for women having urban
residence, non-Muslim religious denomination and husbands being employed as
professionals/businessmen as compared with their counterparts having rural
residence, Muslim religious denomination and husbands being employed in
services other than professional or business, respectively. Controlling for other
variables in the analysis, contraceptive usage increases significantly with
educational levels of women and that of their husbands, current age, duration of
marriage and number of living sons, while it declines with child loss experience.
For each of the interval variables, the squared term has a sign that is opposite the
linear term, implying a decreasing effect, but none of the squared terms are
statistically significant in the final models.
Regarding changes in the effects of the socio-economic, sociocultural and
demographic variables when the ready, willing and able variables are controlled
for in the analysis, possibly the most noteable point is the narrowing down of
regional variations in contraceptive use when the ready, willing and able variables
are controlled in the analysis. This is possibly because the effect of region of
residence on current use is partly transmitted through the ready, willing and able
variables. It should be mentioned that the analysis in the previous section indicated
that region of residence had strong effect on each of those three variables.
However, the regional variations in contraceptive use do not totally diminish when
the ready, willing and able variables are controlled in the analysis. Thus, the ready,
willing and able variables do not entirely account for the regional variations in
contraceptive use. This unexplained variation could be owing to several reasons.
First, the creation of Sylhet division, which isolates the sections of the former
Chittagong division that have the lowest use rates, results in wider divisional
differences than existed previously (Mitra and others, 1997:55). Furthermore, this
unexplained variation could be owing to the different socio-economic status of the
respondents across the regions. In particular, the socio-economic scenario of
Sylhet division is different from the rest of the country in several aspects including
because it has a larger percentage of affluent expatriate and semi-expatriate
population which may not have similar views regarding fertility control compared
to the much larger non-expatriate population of the country. In this study,
socio-economic status of women were partly controlled for. However, further
studies need to be undertaken to understand the reasons for lower use rate among
58
Asia-Pacific Population Journal, Vol. 21, No. 1
women of Sylhet division in particular. Such an understanding may help to provide
appropriate services and possibly bring the Sylhet division in line with the rest of
the country in terms of contraceptive use. The magnitude of the effect of number of
living children on current use reduces drastically and diminishes when the ready,
willing and able variables are introduced in the final model. This is probably
because the effect of number of living children on current use is largely transmitted
through the ready variable as was found in the previous section. Although infant
and child mortality has been declining in Bangladesh, it is still high by international
standards (Mitra and others, 1997). In this study, child loss experience has been
found to have a strong negative significant effect on current use regardless of the
presence of the ready, willing and able variables. When the effect of child loss
experience on readiness, willingness and ability to use contraception was
examined, it was found to have negative effects on all of them, however, the effect
was statistically significant only for willingness. The results thus seem to suggest
that the effect of infant and child mortality on contraceptive use is only partly
transmitted through willingness to use contraception and because of this, the
magnitude of its effect on contraceptive use does not change much when the three
variables are controlled in the analysis. Similarly, the effects of current age and
marital duration on contraceptive use remain strong after controlling for the three
variables. When the effects of current age and marital duration on readiness,
willingness and ability were examined, they were found to be statistically
significant only for ability and readiness, respectively.
Table 4. Logistic regression for current use of contraception, Bangladesh,
1996-1997
Variables
Easterlin/Coale variables
Readiness (Not ready)
Ready
Willingness (Not willing)
Willing
Ability (Not able)
Able
Model A
Coefficient
Model B
Coefficient
Model C
Coefficient
-2.199*
-2.002*
-----
-2.126*
-2.029*
2.396*
--
2.111*
Residence/regional variables
Place of residence (Urban)
Rural
Region of residence (Sylhet)
Barisal
Chittagong
------
--0.502*
-1.247*
0.668*
--0.479*
-0.920*
0.416*
.../
Asia-Pacific Population Journal, April 2006
59
Table 4. (Continued)
Variables
Dhaka
Khulna
Rajshahi
Sociocultural variables
Religion (Muslim)
Non-Muslim
Socio-economic variables
Women's education (No education)
Primary incomplete
Primary complete and above
Current work status (Not working)
Working
Husband's occupation(Prof./business)
Others
Husband's education (No education)
Primary incomplete
Primary complete
Secondary and above
Demographic variables
Current age
Current age squared
Number of living children
Number of living children squared
Duration of marriage
Duration of marriage squared
Number of living sons
Number of living sons squared
Child loss experience
Child loss experience squared
Constant
Model chi-square
Degrees of freedom
Sample size
Model A
Coefficient
Model B
Coefficient
Model C
Coefficient
----
1.300*
1.863*
1.799*
---------------
-0.421*
--0.288*
0.494*
-0.181*
--0.265*
--0.084*
0.050
0.271*
-0.427*
--0.261*
0.375*
-0.136***
--0.301*
--0.146*
0.009
0.163
0.161*
-0.002*
0.476*
-0.055*
0.015
0.000
0.350*
-0.051*
-0.173**
-0.001
-4.103*
0.192*
-0.003*
0.022
-0.011
0.0528***
-0.000
0.1918**
-0.023
-0.167**
0.003
-9.345*
-----------6.010*
1,639.319
1,257.900
0.985*
1.527*
1.365*
2,293.835
3
24
27
7,477
7,362
7,343
Note: The reference category for a categorical variable is in the parentheses.
* indicates p<0.001, ** indicates p<0.01; and *** indicates p<0.05.
60
Asia-Pacific Population Journal, Vol. 21, No. 1
Conclusion
The present study attempted to measure motivation or readiness, willingness
and ability to regulate fertility and examine their impacts on fertility regulating
behaviour of women in view of the rapid fertility decline in Bangladesh. All
analyses in the present study were limited to fecund and non-pregnant women. The
“ready” precondition has received considerable attention in the economic literature
while the “able” precondition has been extensively examined in the family
planning literature. By contrast, “willingness” has not received ample attention
mainly because it is harder to measure. The results of the present study indicate that
a vast majority of women satisfy those three conditions. With the exception of
women without living children, most women want to control their reproduction.
Most women consider fertility regulation as legitimate and acceptable on
normative and health related grounds, that is, they are willing to adopt fertility
regulation, counter to established traditional beliefs, codes of conduct, moral
objections and health concerns. This is an important result as none of the previous
studies in Bangladesh including the one by Ahmed (1987) examined all issues
related to willingness to regulate fertility. In addition, it was found that for most
women family planning methods are available, accessible and convenient in terms
of costs. Thus, this study has established that most women now want to control
their fertility and consider fertility control as normatively acceptable, as well as
convenient in terms of availability/accessibility and cost. In other words, the
sociocultural changes which are favourable to fertility transition have already
taken place in Bangladesh. The results confirmed the greater importance of cultural
factors to willingness, and socio-economic and demographic factors to readiness to
use contraception. It was also found that while they are willing and able, the
majority of women with no living children are not ready to use contraception. This
is also partly why more women are willing to control fertility than they are ready.
The results of multivariate analysis suggested that readiness and willingness
show up as independent factors to ability to practice and therefore they may have
been part of the fertility transition at some point. However, the authors are unable
to suggest at what point of the transition readiness and willingness directly played
roles in the fertility decline. The fact that the analysis shows ability, willingness,
readiness to be strongly associated with contraceptive use, at a time when more
than 50 per cent of eligible women were using contraception, and had been for
some time, does not constitute a full test of Easterlin’s theory since one does not
know which of the changes occurred when. It is possible that those changes in
attitudes occurred after changes in fertility control. To explore this further, the
Matlab longitudinal data can be used although the results will not be entirely
Asia-Pacific Population Journal, April 2006
61
nationally representative. The present study seems to suggest that fertility
transition is well underway in Bangladesh as a vast majority of eligible women
satisfy the three preconditions of fertility decline. Looking back at the national
estimates for Total Fertility Rate (TFR) imply that the rapid decline in TFR from
around five to around three first occurred in 1993-1994 based on the Bangladesh
Demographic and Health Survey (BDHS) and after that it has remained almost
stable. To explore whether in addition to ability, readiness or willingness or both
have also played roles in the fertility transition, one needs at least to measure those
variables based on national population surveys conducted during the pre-transition
and post-transition periods. However, lack of BDHS prior to the major fertility
decline mentioned above somewhat restricts such analysis. As explained earlier,
given the limitation of data, the authors particularly overestimated the percentage
who were willing and able and thus the impact of those variables on contraceptive
use are likely to be overestimated in this study. However, this is the best one could
do with nationally representative data from Bangladesh. Also, from the point of
view of study design ideally the pre-conditions should be measured prior to the
measures of fertility control, and ideally, prior to the onset of fertility change and
this would require longitudinal data. But, in this study cross-sectional data are used
so that preconditions and fertility are measured simultaneously. Unfortunately,
there are no such longitudinal data at the national level in Bangladesh.
Alternatively, one can use the Matlab data collected by the ICDDR,B for
performing this analysis. However, results based on such data will not be
representative of the entire country. Finally, contraceptive use has not increased
much in Bangladesh following the last major fertility transition (decline) and
therefore we do not expect that the effects of the three variables on contraceptive
use have changed dramatically since the 1996-1997 BDHS. It is not expected that
using data from more recent surveys would give substantially different results from
those obtained in this study.
62
Asia-Pacific Population Journal, Vol. 21, No. 1
References
Ahmed, B. (1987). Determinants of contraceptive use in rural Bangladesh: the demand for children,
supply of children, and costs of fertility regulation, Demography, vol. 3, No. 24, pp. 361-373
Bernhardt, M.H. and M.M. Uddin (1990). Islam and family planning practices in Bangladesh, Studies in
Family Planning, vol. 5, No. 21, pp. 287-292.
Caldwell, J.C. and others (1999). The Bangladesh fertility decline: An interpretation, Population and
Development Review, vol. 1, No. 23, pp. 67-84.
Carty, W.P., N.V. Yinger and A. Rosov (1993). Success in a challenging environment: Fertility decline
in Bangladesh, Washington DC: Population Reference Bureau.
Cleland, J.C., J.F. Phillips, S. Amin and G.M. Kamal (1994). The determinants of reproductive change
in Bangladesh: Success in a challenging environment, Washington DC: The World Bank.
Cleland, J.C. (1985). Marital fertility decline in developing countries: Theories and the evidence: In
Cleland, John and J. Hobcraft, eds., Reproductive change in developing countries (Oxford,
Oxford University Press).
Coale, A.J. (1973). The demographic transition reconsidered. Proceedings of the International
Population Conference: 53-72. Liege, Belgium. International Union for the Scientific Study
of Population.
Easterlin, R.A. (1975). An economic fertility framework for analysis, Studies in Family Planning, vol.
6, No. 3, pp. 54-63.
Freedman, R.S., S. Khoo and B. Supraptilah (1981). Modern contraceptive use in Indonesia: A
challenge from conventional wisdom. WFS Scientific Reports, No.20. Voorburg, The
Netherlands: International Statistical Institute.
Islam, M.M., M.K. Barua and R. Bairagi (2003). Quality Characteristics of Field Workers and
Contraceptive Use Dynamics: Lessons from Matlab, Bangladesh, Asia-Pacific Population
Journal, vol. 18, No.1, pp. 43-62.
Kamal, N. and A. Slogget (1993). The influence of religiosity, mobility and decision making on
contraceptive use. In Secondary Analysis of the Bangladesh Fertility Survey Data 1989.
National Institute of Population Research and Training (NIPORT), Dhaka, Bangladesh.
Khan, H.T. and R. Raeside (1998). The Determinants of First and Subsequent Births in Urban and Rural
Areas of Bangladesh, Asia-Pacific Population Journal, vol. 13, No. 2, pp. 39-72.
Khan, M.A. and M. Rahman (1997). Determinants of Contraceptive Method Choice in Rural
Bangladesh, Asia-Pacific Population Journal, vol. 12, No. 3, pp. 65-82.
Asia-Pacific Population Journal, April 2006
63
Koenig, M.A. (2003). The impact of quality of care on contraceptive use: Evidence from longitudinal
data from rural Bangladesh, Johns Hopkins Bloomberg School of Public Health,
Department of Population and Family Health Sciences, Baltimore, Maryland, USA.
Knodel, J. (1974). The decline of fertility in Germany, 1871-1939 (Princeton, Princeton University
Press).
Lesthaeghe, R. and C. Vanderhoeft (1998). Ready, willing and able: A conceptualization of transitions
to new behavioural forms. Paper presented for the National Academy of Sciences meeting
on the “Social Dynamics of Fertility Change in Developing Countries”, Washington DC,
USA.
Lesthaeghe, R. (1977). The decline of Belgian fertility, 1800-1970 (Princeton, Princeton University
Press).
Mannan, H.R. (2002). Factors in contraceptive method choice in Bangladesh: Goals, competence,
evaluation and access, Contraception, vol. 65, No. 5, pp. 357-364.
Mitra, S.N., A. Al-Sabir, A.R. Cross and K. Jamil (1997). Bangladesh Demographic and Health Survey,
1996-97. Calverton, Maryland and Dhaka, Bangladesh: National Institute of Population
Research and Training (NIPORT), Mitra and Associates, and Macro International Inc.
National Institute of Population Research and Training (NIPORT) (2001). Bangladesh Demographic
and Health Survey, 1999-2000. Macro International Inc., Calverton, Maryland and Mitra
and Associates, Dhaka, pp. 45-76.
Rosero-Bixby, L. and J.B. Casterline (1993). Modelling diffusion effects in fertility transition,
Population Studies, vol. 47, No. 1, pp. 147-167.
Stevens, W.K. (1994). Poor land’s success in cutting birth rate upsets old theories. In Griffiths, R.J.
(ed.), Developing World 95/96, Sixth Edition: 201-202, Guilford: The Pushkin Publishing
Group.
Tsui, A.O., D.P. Hogan, J.D. Teachman and C. Welti-Chanes (1981). Community availability of
contraceptives and family limitation, Demography, No. 18, pp. 615-625.
Ullah, M.S. and N. Chakraborty (1993). Factors Affecting the Use of Contraception in Bangladesh: A
Multivariate Analysis, Asia-Pacific Population Journal, vol. 8, No. 3, pp. 19-30.
64
Asia-Pacific Population Journal, Vol. 21, No. 1
Singapore’s Family Values: Do
They Explain Low Fertility?
Most Singaporeans identified positively with the nuclear
family structure and “standard” family roles. Also, Singaporeans are
generally pro-children. In this context, it is difficult to support those
views that argue that Singaporeans have become
highly individualistic and hedonistic.
By Alexius A. Pereira*
To the Government of Singapore, the country’s declining marriage and
fertility rates are serious national problems. It believes that those trends will have
negative consequences for economic growth and Singapore’s overall quality of life
in the future as Singapore faces a “greying population”. In 2003, there were 21,962
marriages registered, lower than 2002 (23,189), the 1990s (average 24,000) and
the 1980s (average 23,000) (Singapore Department of Statistics, 2004: 14).
Between 1970 and 1975, Singapore’s total fertility rate averaged 2.6; in 1980, it
was 1.80; in 1986, 1.43; in 1990, 1.83; in 2000, 1.60; and in 2003, it had fallen to
1.24.1 During the same period, the population census also found that there was a
* Assistant Professor, Department of Sociology, National University of Singapore; e-mail:
[email protected].
Asia-Pacific Population Journal, April 2006
65
higher proportion of Singaporeans remaining unmarried. In the Singapore Census
of Population 2000, for the age group 30-34, one in three Singaporean males and
one in five Singaporean females were unmarried (Singapore Department of
Statistics, 2001: 2). The State is particularly concerned that Singapore’s future
economy will be unable to sustain an ageing population, where 20 per cent of the
population would be aged 65 and older by 2030 (Singapore Department of
Statistics, 2002: 6).
The State has implemented a wide variety of measures over the past 20 years
in an attempt to reverse the declining marriage and fertility rates. Those measures
have included fiscal incentives as well as ideological persuasion (mainly through
campaigns such as the “Romancing Singapore” festival). However, by 2004,
official statistics showed that the fertility rate had fallen even further, and that there
were even more “singles” in Singapore than ever. This had led one prominent
Singaporean statistician (Paul Cheung, Chief Statistician, Singapore Department
of Statistics, 1983-2004) to opine:
In the 1980s, many singles were single by circumstance. If they did
get married, they would probably have two or three kids. So the
Social Development Unit came in and stabilised things, and the
birth rate actually rose for a few years… But now [2004], more
people choose to stay single. And couples choose to have one or no
kids. Their lifestyle choices have changed. So influencing the birth
rate now will require different methods. (Singapore Straits Times,
23 May 2004)
In other words, while it might have been true that Singaporeans in the recent
past agonized over their inability to get married and have children, the new view
argues that most contemporary Singaporeans intentionally do not want to get married
or to have children. Many policymakers believe that this new view on the family is
owing to changes in Singapore’s family values. The Government’s view is clearly
laid out in the executive summary of Family Matters, a State-commissioned report of
the Public Education Committee on Family:2
1. Values guide the decisions that we make: our relationships, our work and life
as a whole; as well as the responsibilities that come with them. Just as
families are the basic building blocks of the society, values are the
foundations that underpin the family. Family values are the set of tenets
necessary for holding a family together. The emphasis given to teaching
values in schools and the promotion of Singapore Family Values underscore
their importance.
66
Asia-Pacific Population Journal, Vol. 21, No. 1
2. Singapore society has seen tremendous changes in the past few decades.
Globalisation (sic), technological change and the Internet have
expanded our spheres of influence beyond our immediate environment.
Work and family have become highly interdependent with the rise of
dual-income families. Parents face the “time-bind”, which often results
in inadequate value transmission to their children. These challenges
have the potential to erode the values that ensure the well-being of
families.
3. In recent years we have seen a gradual shift in long-held attitudes towards
relationships, marriage and family. More Singaporeans are remaining
single, delaying marriage and having fewer children. Many place priorities
on careers and other life goals, while holding high but often unrealistic
expectations about their life partners. Youth are adopting increasingly
liberal views towards sexual intimacy, marriage commitment,
childbearing, etc. Efforts must be expended now to foster positive attitudes
and strengthen our social institutions. (PECF, 2002: 11)
In response to this, the Government of Singapore implemented a broad range
of measures to reinforce the family institution as a key strategy to reverse the
declining marriage and fertility rates. In his first National Day Rally Speech, the
country’s third Prime Minister Lee Hsien Loong promised to make “Singapore: A
Great Place for Families”, which has become the tagline for the new “pro-family”
policies (see http://www.family.gov.sg/). In addition to even more fiscal incentives
for procreation, the State has taken the lead in completely restructuring the civil
service to create a “pro-family” environment, beginning with the implementation
of a “five-day working week”, with weekends set aside for “family time” (see
http://aboutfamilylife.org.sg/). The state also feels very strongly that family values
must be “strengthened”, introducing a variety of programmes to promote marriage
and childbearing.
Research question
It is fairly clear that several influential individuals and senior policymakers in
Singapore believe that the society’s family values have been eroded, and this
erosion has been a key factor in causing the decline in marriage and fertility rates.
The cause of the erosion, according to the State, is “economic development”, and
the more recent process of globalization. This study will therefore focus on
examining Singapore’s family values at the turn of the millennium. More
specifically, it intends to analyse and explain whether Singaporeans actually value
the family (as an institution), marriage, parenthood, motherhood, childbearing, and
Asia-Pacific Population Journal, April 2006
67
other issues regarding the family. It posits that if Singaporeans hold pro-family
views, it can be concluded that Singapore’s family values are strong, and vice
versa. Towards this end, this study therefore intends to identify and account for
Singapore’s family values at the beginning of the new millennium. Since the
Government of Singapore assumes that “younger” Singaporeans appear to be
facing a greater risk of value erosion (as a consequence of modernization,
industrialization or globalization), it is also worth examining whether there is a
difference in value system between “younger” and “older” Singaporeans.
Theory of intergenerational value change
It is worth noting that the position adopted by the contemporary Government
is one where it feels that “value change” seems to be the primary reason behind
fertility decline. As sociologists have long noted, there could possibly be many
different reasons behind fertility decline (see Van Krieken, 1997 for a summary).
Indeed, most sociologists would argue that fertility decline arises from a
combination of factors, some sociological, others economic, political and even
historical (Alter, 1992; Gillis, 1996). Hence, it is worthwhile to examine the
Government of Singapore’s logic of “value change”, which seems to be drawn
from existing theories of “intergenerational value change”. The central argument
of those theories, which originated from various strands of modernization theory,
is that the processes of industrialization and economic growth will lead to greater
“individualization” in society. According to Beck:
Traditional bonds (kin, clan, community) tear apart as industrial
and postindustrial society emerges. The feasibility of living a more
or less detached life is aided by the modern state (1992: 32).
Similarly, Beck-Gernsheim (2002) holds that family life today is characterized by
the “post-familial family”, where the “traditional” family a lifelong officially
legitimated community of father-mother-child, held together through emotion and
intimacy is being replaced by a diverse array of lifestyles. She explains that
“individualization”, also brought about by changes in modern social institutions
such as the State, is the key driver behind this mindset change; the result is that
people now think and act as individuals rather than based on strong kinship ties and
family obligations (Beck-Gernsheim, 2002: 41).
Inglehart (and associates) holds a similar view on value change. With
industrialization and economic growth, people place “less emphasis on traditional
cultural norms ... especially those norms that limit self-expression” (Inglehart
1997:33-35). In this sense, Beck and Inglehart indirectly agree that the family is
less socially crucial than it once was, as it is no longer the key economic or
68
Asia-Pacific Population Journal, Vol. 21, No. 1
socialization unit in modern (and post-modern) societies. As Zimmerman (2001)
argues, changes in the economy, which in turn lead to changes in the demographic
structure of society and changing roles of women, lead to people valuing the family
differently. The rise of capitalism and the Industrial Revolution brought challenges
to the social norms that “…buttress traditional family values and patriarchal norms
of male dominance” (Inglehart and Norris, 2003:16). As more women enter the
labour force, with increasing education and qualifications and the emergence of
feminist movements, gender roles will have to change and adapt to the new
economic structure. Thus, modernization theorists suggest that social trends like
late marriage, late parenthood, smaller family units, the “double-income-no-kids”
(DINKs) syndrome and increase in divorce rates are becoming more common
nowadays and can be attributed to the process of “individualization”. In particular,
the authors believe the “affluence effect” is the most pronounced factor behind
individualization, as affluence allows individuals and family “… to experiment
with different forms of self-expression and individuality ... sweeping aside
traditional values rooted in generations of want and scarcity” (Zimmerman,
2001:74). As such, with individualization taking hold, modernization theorists
would argue that the notion of having “obligations” would eventually erode.
Previously, “obligations” were central to social life in traditional societies, as
individuals were necessarily held by their obligations towards family, religion or
political authorities. With greater individualism, people now have the power of
“choice” over what they do. The final link between value change and declining
fertility is therefore the belief that marriage, starting a family and childbearing are
now personal choices rather than social obligations.
The theory of intergenerational value thus posits that individualization would
be more pronounced in the generational group that experienced industrialization
and economic growth directly. With greater propensity towards individualization,
the generation in question would be more likely to feel that the family (marriage,
family structure and roles) and childbearing are less important in their lives.
Although those views have been criticized as being “too monochrome and too
one-dimensional” (Smart and Shipman, 2004: 506), there is no doubting that
modernization theories do have “ideological appeal”, especially to the so-called
“conservative segments of society” (Gillis, 1996; Zimmerman, 2001). This is
because values are used as guides for behaviours, as they tell people what they
ought or ought not to do; they “incorporate ideas, symbols, and beliefs that help
people make sense of their lives and the world” (Zimmerman, 2001:65-6). Thus,
family values are supported by norms, rules and laws that act as moral compasses
to “help” people realize the “ideal family” type. Family values are thus conceptions
of what is desirable and looking after the family well-being represents the goal of
Asia-Pacific Population Journal, April 2006
69
family policy. Clearly, in such a discourse, the fingerprints of certain interest
groups which might be the State, or segments of the State captured by interest
coalitions are evident. In such a scenario, the notion of “value change” often tends
to imply a “change for the worse”. Hence, some sort of intervention (or policy) is
necessary, as seen in the case of the Government’s pro-family policies.
This paper will examine Singapore’s family values, which are defined as how
individuals value the institution of the family, rather than “what family life ought to
be”. More specifically, it hopes to analyse whether the Singapore’s family values
are “strong” or whether the opposite situation of “individualization” has taken
place. If “individualization” has taken place, it could therefore be proposed as a
primary explanation for Singapore’s declining marriage and fertility rates.
Methodology
The study is based on an analysis of the Singapore-leg of the World Values
Survey, conducted in 2002. According to the International Network of Social
Scientists, the organization in charge of the survey:
The World Values Survey is a worldwide investigation of
sociocultural and political change. It is conducted by a network of
social scientist at leading universities all around world. The survey
is performed on nationally representative samples in almost 80
societies on all six inhabited continents. A total of four waves have
been carried since 1981 (http://www.worldvaluessurvey.org/).
An abridged version of the survey was conducted in Singapore by a team from
the Department of Sociology, National University of Singapore.3 The
WVS-Singapore 2002 was constructed to emulate the proportions of major social
categories of the Singapore population, including by gender and ethnicity (see table
1). This dataset consisted of views by Singaporeans on various aspects of the family.
Table 1. WVS-Singapore 2002 Sample, by gender and ethnic group
Indian
Chinese
Ethnicity
(Percentage)
Male
Female
Total
39
53
92
6.1
603
602
1,205
79.9
Malay
98
104
202
13.4
Others
6
7
13
0.9
Total
746
766
1,512
100
Gender (percentage)
49.3
50.7
100
70
Asia-Pacific Population Journal, Vol. 21, No. 1
This study was interested in examining whether Singapore actually “suffered”
from weak family values. As mentioned earlier, “weak family values” would
refer to individuals placing less importance on various aspects of the family,
such as getting married or bearing children. Weak family values can also be
understood as implying the presence of “individualization” among the
population. Therefore, the first area of inquiry would be to identify how
Singaporeans value the family “as an institution”. Do Singaporeans think the
family is important? How important is it when compared against other
institutions and social groups (such as friends and work colleagues)? The
second area of inquiry would involve identifying how Singaporeans value
marriage as an institution. Do Singaporeans think the family is an outdated
institution? How do Singaporeans feel about women who want to be unmarried
single parents by choice? The third area of inquiry would involve identifying
how Singaporeans value the nuclear family, as a structure, as well as
parenthood. What do Singaporeans think is the “ideal” structure for the
family? Must it involve both a husband and a wife? What is the “ideal” number
of children a family ought to have, if any? Can career women be as effective in
bringing up children? Is being a housewife considered fulfilling?
In addition, since the basic line of inquiry seeks to test the theory of
intergenerational value change, in lieu of applicable comparative data, this
study artificially segments the sample into two generational cohorts (younger
generation and older generation). If there are differences in views between the
two cohorts, it might indicate that there are generational differences and
therefore possibly a change in values across generations. This however cannot
be taken as definite proof of generational change, as only a time-series
comparison would qualify. In addition, there are further limitations to this
study, which seeks aggregate indicators through the homogenization of the
sample. This is because Singapore (and therefore the WVS sample) is highly
heterogeneous, in terms of ethnicity, religious affiliation, as well as in terms of
class and educational differences. Yet, since this is a preliminary study, it is
more important to tease out wider aggregate trends first, on order to get a sense
of what Singaporeans feel before more detailed disaggregated analysis
follows. Thus, the only social variable examined is that of gender, as the study
is interested in finding out whether there were any significant differences
between the views of men and women. Gender differences are possible
because the issue of the family is intrinsically linked to gender and gender
roles within the family.
Asia-Pacific Population Journal, April 2006
71
Data
The family
At the aggregated societal level, data generated from the WVS-Singapore
2002 gave very clear indicators about the family values held by Singaporeans.4
Firstly, over 91 per cent of the respondents indicated that they felt that the family
was “very important” (see table 2). However, it is significant to note that this figure
is slightly lower than the mean of 31 countries (93.2 per cent indicated that family
was “very important”) in the WVS5 (Fourth Wave) conducted between 2000 and
2001. Interestingly, the Singaporean aggregate response was lower than the
response from the United States of America (95.3 per cent), South Africa (95.7 per
cent) and Nigeria (98.9 per cent), but higher than China (60.2 per cent), and the
Republic of Korea (89.6 per cent) (WVS, 2000).
Table 2. Views on “the family” as an institution (percentage)
Male
Female
Combined
90.8
92.7
91.8
Rather important
8.7
6.7
7.7
Not very important
0.5
0.5
0.5
Very important
Not at all important
Total
0.0
100.0
0.0
100.0*
0.0
100.0
Note: Figure due to rounding.
When compared to other social institutions, the “family” was ranked as being the
most important (see table 3).
Table 3. Percentage of respondents indicating that these social aspects are
“very important” (percentage)
Male
Female
Combined
Family
90.8
92.7
91.8
Work
59.4
45.5
52.3
Friends
41.7
37.4
39.5
Religion
33.5
38.3
35.9
Leisure time
27.4
27.9
26.3
Politics
10.9
9.0
10.0
Singaporeans also generally felt that “there ought to be more emphasis placed on
family life”, especially when compared to other aspects of social life (see table 4).
72
Asia-Pacific Population Journal, Vol. 21, No. 1
Table 4. Priorities in social life (percentage)
Combined (male/female)
More emphasis on family life
Good
Don't mind
Bad
92.8
(91.7/93.8)
4.6
(5.1/4.0)
2.3
(3.0/1.6)
66.8
28.6
4.1
(70.5/63.3)
(26.3/30.8)
(2.7/5.4)
52.1
40.3
7.1
(53.2/51.1)
(39.0/41.7)
(7.6/6.7)
Less emphasis on money and material possessions
37.9
(35.8/39.9)
45.9
(47.6/44.3)
16.1
(16.6/15.7)
Less importance placed on work
28.6
(27.6/29.7)
37.4
(37.3/37.4)
33.7
(34.8/32.6)
More emphasis on the development of technology
Greater respect for authority
Singaporeans also appeared to prioritize social contact with family members.
Nearly three quarters of the sample indicated that they “frequently” spent time with
family members, whereas only half indicated that they frequently spent time with
friends (see table 5).
Table 5. Social contact - at least weekly (percentage)
Male
Female
Spend time with parents or other relatives
72.7
76.3
74.5
Spend time with friends
59.9
43.8
51.8
Spend time socially with colleagues from
work or profession
30.0
23.0
26.4
19.0
18.2
18.6
13.3
6.3
9.7
Spend time with people at place of worship or
religious organization
Spend time socially with people at sports clubs,
voluntary or service organization
Combined
Based on those responses, it can be concluded that Singaporeans appear to
value the family, as an institution, very highly. They also view various aspects of
family life, such as contact with family members, as being important in their lives.
Asia-Pacific Population Journal, April 2006
73
Marriage
More than three quarters of the respondents think that marriage is not an
outdated institution (see table 6). However, over 70 per cent of respondents who
were single (at the time of the survey) disagreed that “marriage was an outdated
institution”.
Table 6. Views on “marriage is an outdated institution” by marital status
(percentage)
Combined
(male/female)
Marital status
Married
Divorced
Separated
Widowed
Single
Percentage of sample
Disagree
45.1
83.2
(40.7/49.5)
(84.2/82.3)
1.5
67.9
(1.2/1.7)
(77.8/61.5)
0.5
65.8
(0.7/0.3)
(60.0/100.0)
1.6
98.9
(0.4/2.7)
(100.0/100.0)
51.2
72.6
(56.8/45.8)
(71.9/73.4)
Total: 100
Mean: 77.8
(49.3/50.7)
(77.1/78.4)
In addition, when the respondents were asked: “If someone says a child needs
a home with both a father and a mother to grow up happily, would you tend to agree
or disagree?”, over 93 per cent (standard deviation of 0.512) indicated that they
agreed. This implies that Singaporeans valued the “nuclear” family structure,
which consists of a husband and a wife, along with their children, living in a
household. Further, most Singaporeans disapproved of women as single parents by
choice (defined as a woman choosing to have children without having a stable
relationship with a man) (see table 7).
74
Asia-Pacific Population Journal, Vol. 21, No. 1
Table 7. Views on “family structure" - agree (percentage)
Male
Female
Combined
Nuclear family structure important
95.3
91.6
93.4
Single-parenthood for women by
choice is wrong
74.1
68.1
71.1
The WVS-Singapore 2002 also found that Singaporeans tended to hold
rather “modern” views towards family roles (see table 8). Over 70 per cent of
the respondents felt that a working mother could establish just as warm a
relationship with their children, than a mother that does not work (outside the
home). In addition, over 80 per cent of respondents felt that both the husband
and the wife should contribute to the household income. Finally, nearly two
thirds of the respondents indicated that being a “housewife” was just as
fulfilling as working for pay.
Table 8. Views on “family roles” - agree (percentage)
View
Male
Female
Combined
Working mother alright
67.5
73.4
70.5
Housewife fulfilling
68.6
68.6
68.6
Both spouses should contribute to
household income
77.8
82.5
80.2
Interestingly, those views on “family roles” saw some variation between the
male and female responses. For instance, a slightly larger proportion of women
than men felt that “a working mother could establish just as warm a relationship
with their children”, and also the view that “both spouses should contribute to the
household income”. Overall, the views on family roles can generally be understood
as being fairly “modern” views, as opposed to “traditionally conservative” views,
which would disapprove of working women. By contrast, the views indicated that
most Singaporeans felt that women could adopt both roles (career woman or
housewife), and that it was entirely the woman’s own choice, as opposed to only
being allowed to perform traditional female roles (mother, wife and homemaker).
This therefore indicates that employment, by itself, is not viewed as being an
obstacle to getting married as well as having or raising children.
Asia-Pacific Population Journal, April 2006
75
Childbearing
On the issue of having children, very few Singaporeans indicated that they
did not wish to have children (1.3 per cent of the sample), whereas over 80 per cent
of the sample felt that the “ideal size of the family” included having either two or
three children (see table 9).
Table 9. Views on “ideal size of family” (percentage) by marital status
Number of children
(male/female)
Combined
(male/female)
Total
Married
(n = 683)
Single
(n = 829)
(n = 1,512)
None
0.9
(0.6/1.1)
1.6
(1.7/1.4)
1.3
(1.3/1.2)
One
2.5
(3.5/1.7)
3.6
(4.3/2.7)
3.1
(4.0/2.2)
Two
45.0
(42.3/47.2)
53.9
(56.4/51.0)
49.9
(50.7/49.1)
Three
33.9
(34.9/33.0)
28.0
(27.7/28.4)
30.7
(30.6/30.7)
Four
14.1
(14.1/14.1)
9.1
(6.2/12.4)
11.3
(9.4/13.2)
2.4
(3.1/1.7)
3.9
(3.7/4.1)
2.2
(2.1/2.3)
Mean
2.74
2.55
2.65
Median
3.00
2.00
2.00
2
2
2
Standard Deviation
1.103
1.146
1.162
Variance
1.217
1.314
1.349
Five or more
Mode
It was also interesting to note that there were only minor differences in views
between those who were married at the time of the survey and those who were
single. In aggregate terms, it could be argued that those that were single generally
preferred having slightly fewer children than those that were married. However,
76
Asia-Pacific Population Journal, Vol. 21, No. 1
most singles did not indicate that they did not want children at all. When comparing
the responses of the “ideal” number of children with the actual number of children
the respondents had, there were some variances (see table 10).
Table 10. Ideal number of children, for married respondents (n = 683)
Combined
(male/female)
Number of children
Actual
Ideal
None
12.3
(15.6/9.6)
0.9
(0.6/1.1)
One
19.9
(20.1/19.6)
2.5
(3.5/1.7)
Two
36.2
(37.1/35.6)
45.0
(42.3/47.2)
Three
22.5
(16.9/27.0)
33.9
(34.9/33.0)
Four
4.8
(7.1/3.1)
14.1
(14.1/14.1)
Five or more
4.3
(3.2/5.1)
3.6
(4.5/3.0)
100
100
Total
There could be some possible explanations for this outcome. First, the
desired number of children can be explained as being higher than the actual number
because some families may continue to have children (i.e., in the future). The
second possible reason is that Singaporeans would like to have more children than
they currently have, but choose not to do so, for various reasons. In this sense, the
second reason could be a policy concern.
On a related issue, only slightly more than half the respondents felt that
childbearing, by itself, is not necessary for a woman to feel “fulfilled” (54.5 per
cent agreed to the statement “Childbearing is necessary for a woman to feel
fulfilled”) (see table 11). This would suggest that nearly half of the sample were of
the view that having children was more a personal choice than a social obligation.
While it might follow that respondents having indicated that women needed to
have children to feel fulfilled will perceive that it is important to have children, it is
more significant that most of those that indicated that having children is a personal
Asia-Pacific Population Journal, April 2006
77
choice still stated that they valued having children. Thus, it could be concluded that
most Singaporeans valued having children.
Table 11. Views on childbearing (percentage)
Combined
(male/female)
Childbearing is necessary for
women to feel fulfilled
Standard deviation
Necessary
Not necessary
Don’t know
54.5
(52.3/56.6)
42.7
(43.1/42.2)
2.9
(4.6/1.2)
1.354
Up to this point, the emergent data suggest that there was not too much
variation between the views of men and women in the survey. As mentioned
earlier, the most significant difference of opinion was found with issues
concerning gender roles within the family.6
Generational change?
While the aggregate data from the Singapore-leg of the World Values Survey
suggest that Singaporeans generally value the family, childbearing and marriage, it
is important to examine whether there is any difference in opinions between age
groups. Age groups are important to this analysis because the Government of
Singapore and various other state agencies seem to think that so-called “younger”
Singaporeans appear to face a higher risk of value erosion or “individualization”,
as suggested in the PECF’s recommendations (article 3, as mentioned earlier). For
this preliminary study, “younger” Singaporeans are defined as being born after
Singapore’s independence in 1965 (i.e. those who are aged 37 years old and
under), while those born before independence (i.e. those who are aged 38 years old
and over) would be considered “older” Singaporeans.7 The year of independence
as the dividing point was chosen as it was assumed that each group would have
been growing up and socialized under different circumstances. Those born before
independence would probably have faced Singapore’s earlier economic hardships,
whereas those born after independence were probably growing up during
Singapore’s economic boom. If there was no difference of opinion between the two
age groups, then it could be suggested that there was no value change across time.
Although there is a statistical problem in categorizing the age groups as the
Singapore-leg of the WVS included 978 respondents aged 37 years old and under
during the year 2002, while there were 532 respondents who were aged 40 years
old and over the resultant data are still useful in shedding some light on that issue.
78
Asia-Pacific Population Journal, Vol. 21, No. 1
The oldest respondent was born in 1922, making him around 80 years old at the
time of the survey, there were 83 respondents who were 15 years old (the minimum
age to participate in the survey). The mean age of all the respondents was between
32 and 33 years old.
For this analysis, the responses by each age group to several key issues were
compared, and then ranked on the basis of the degree of difference of opinion. Just
by comparing means, it was interesting that there was almost no difference in the
two age groups’ opinion that “a working mother can establish a warm relationship
with her children”, but by contrast, there was a very significant difference of
opinion as regards “childbearing is necessary for women to feel fulfilled” (see
table 12).
Table 12. Views on family life, by age cohorts and gender (percentage)
Younger
Men
n
Women
Older
Total
Men
Women
Total
516
462
978
229
303
532
Agree that childbearing is necessary
for women to feel fulfilled
43.4
46.9
45.0
72.5
71.6
72.0
Agree that marriage is an outdated
institution
25.0
21.6
23.4
12.2
17.2
15.0
Approve of single parent mother
by choice
18.2
24.1
21.0
11.8
13.8
12.9
Nuclear family important
94.0
90.5
92.3
98.3
93.4
95.5
Agree that being a housewife
can be fulfilling
69.2
65.2
67.4
67.2
73.7
70.9
Family is very important
90.3
92.0
91.1
92.1
93.7
93.0
Agree that a working mother can
establish a warm relationship
with children
70.5
72.0
71.2
60.5
75.5
69.1
Note: “Younger” category: Respondents aged 37 years old and below as of the year 2002; “older”
category: respondents aged 38 years old and above as of the year 2002.
There also appears to be some difference of opinion between the two age
groups over the issue of marriage as an outdated institution. Younger respondents
were willing to agree that marriage was an outdated institution compared to older
respondents, with almost a 10 per cent difference in opinion. Although the overall
number of respondents for both groups is still relatively low, it again suggests that
there is some value change. The implication here is that younger Singaporeans
Asia-Pacific Population Journal, April 2006
79
seem to now accept alternative institutions to marriage, such as cohabitation
(without marriage) and singlehood. The other issue which generated a significant
difference in opinion is that it is acceptable for women to be single parent by choice
(nearly 8 per cent difference). More specifically, younger Singaporeans appear
more open to that possibility, whereas older Singaporeans generally expressed
disapproval at such a personal choice.
It was also interesting to examine the difference in opinion on the issue of the
“ideal number of children in a family” between the age groups (see table 13). In
general, younger Singaporeans indicated that they viewed having two children per
family as being ideal, whereas older Singaporeans were split between two and
three children per family.
Table 13. Ideal number of children in the family, by age cohort and gender
(percentage)
Younger
n
Men
516
Women
462
Older
Total
978
Men
229
Women
303
Total
532
None
1.9
0.9
1.4
0.0
2.0
1.1
One child
3.3
2.4
2.9
5.7
1.7
3.4
Two children
57.2
57.1
57.2
36.2
37.0
36.7
Three children
28.7
27.5
28.1
35.4
35.6
35.5
Four children
5.8
10.2
7.9
17.5
17.8
17.7
Five or more children
1.9
1.1
1.5
3.9
3.7
4.3
Don’t Know\ no answer
1.2
0.9
1.0
1.3
1.3
1.3
Total
100
100
100
100
100
100
Mean
2.48
2.53
2.51
2.91
2.94
2.92
1.096
0.990
1.047
1.276
1.329
1.305
SD
There are several possible explanations for this difference of opinion. On
the one hand, it could be argued that older Singaporeans hold traditional
values, which in the Asian case would refer to valuing large families and
having large numbers of children per family, whereas younger Singaporeans
are more modern in that they prefer smaller close-knit nuclear families, usually
consisting of parents and two children. Alternatively, there is the possibility
that an economic reason is behind the difference, as younger Singaporeans feel
80
Asia-Pacific Population Journal, Vol. 21, No. 1
that contemporary Singapore’s cost of living is very high, and therefore it is
uneconomical to have more than two children per family. Conversely, older
Singaporeans would view the fact that having more children was economically
functional, as the children would jointly contribute to supporting the parents in
their old age. At this stage, it is impossible to pinpoint the actual reason for this
difference in opinion without engaging in deeper qualitative research on the
issue. However, what is clear is that there are some differences of opinion
concerning certain issues, which could be an indicator of value change across
generations. Given that the main differences in views were on the issues of
women’s childbearing being a choice, that marriage is an outdated institution,
and acceptance of single parenthood by choice, this suggests that indeed some
degree of individualization has taken place for the “younger” generation.
Conclusion and policy implications
At an aggregate level, the data from the WVS-Singapore 2002 suggest that
most Singaporeans strongly value the family and marriage as an institution, as well
as family life as being “very important”. Most Singaporeans identified positively
with the nuclear family structure and “standard” family roles. Also, Singaporeans
are generally pro-children. In this context, it is difficult to support those views that
argue that Singaporeans have (already) become highly individualistic and
hedonistic. For example, Singaporeans have not given any indication that they
favour a lifestyle of single-parenthood or unmarried cohabitation. There is also no
indication that Singaporeans solely think of themselves, their work or friends,
ahead of family members.
Thus, at the aggregate level, those views mirror earlier findings from
studies done after the 1990 Singapore Census of Population, which found that
marriage and parenthood were important “personal goals” for the large
majority of Singaporeans (see Quah, 1998 and 1999). This suggests that
Singapore faces a “social problem”, which can be defined as a sizable gap
between the ideals and the reality in society (Coleman, 1998). It is clear that
most Singaporeans value marriage, parenthood and childbearing, but
somehow do not seem able to achieve those personal goals. It could be further
argued that the problem is increasing because the social outcomes today are
even further away from the ideals than in the corresponding period a decade
earlier. In this sense, low marriage and fertility rates are both a national as well
as a personal issue. As a personal issue, it is probably highly likely that most
Singaporeans do feel some degree of anxiety and stress over not being able to
get married or to have children (or as many children as they would like to).
Asia-Pacific Population Journal, April 2006
81
Viewed from a different angle, this research proposes that Singapore’s low
fertility is therefore not because of a high degree of individualism among the
people. This finding probably gives greater impetus to the State’s current
pro-family strategy, which is trying to assist Singaporeans in balancing their work
and personal lives, and coping with the perceived high economic costs of having
and raising children. However, from a policy perspective, it is recommended that
the strategy needs to have greater direct impact. This is because at the moment, the
various policies within this strategy are mostly guidelines and recommendations
rather than enforceable laws. Since the introduction of those policies, the
organizations that have been putting in place the improved “work-life”
arrangements are the State’s own agencies, ministries and statutory boards.
Outside of the mandatory extension of maternity leave (which is upheld by the
law), all the other “pro-family” recommendations remain generally optional for
firms in the private sector. Still, the State’s strategy remains important, as it
demonstrates that the Government is trying to create a pro-family sociocultural
environment. The potential social problem here would be that since economic
issues are deemed to be more important than family life, which might be perceived
as being impossible to achieve, the next generation could completely give up even
trying to form a family. In this sense, the State’s current policy will be useful to
those who would like to form a family but they might view their career aspirations
and other economic issues as potential obstacles.
Viewed from the State’s perspective, there probably is some urgent need to
further promote Singapore as a “pro-family” society. This is because this research
also found that there are already some mindset differences between younger and
older Singaporeans, especially on issues such as childbearing, whether marriage is
outdated as an institution, and being a single parent by choice. The widest
difference of opinion was on the issue of childbearing for women as a personal
choice, where the difference between the cohorts was about 26 per cent. The
difference on all other issues was less than 10 per cent, with four issues returning
less than 5 per cent difference. So while family values appear to be generally
“strong” if the entire sample is viewed as a single cohort, it suggests that the
mindset of individualism has already taken hold among some “younger”
Singaporeans. While the degree of individualism does not appear to be very strong
at the moment, there is the “danger” that this might become the “norm” among
younger Singaporeans in the future. How does the State stem or even reverse this
trend? Given that younger Singaporeans generally demonstrate that they have
“fairly strong” family values, while older Singaporeans have “very strong” family
values, the policy implication is that the State should focus on making Singapore a
pro-family society rather than embark upon an ideological campaign to “improve
82
Asia-Pacific Population Journal, Vol. 21, No. 1
family values”. This is mainly because the latter could have the opposite effect on
what is generally a highly educated and affluent society, which might not view
“orders” from above too favourably. Hence, the State must make hay while the sun
shines; Singapore currently has strong family values, Singaporeans would like to
get married and to have children, and appear to be asking for “help” to do so. If the
State can help those people resolve their personal problems, it is more likely that
they will retain and transmit strong family values to the next generation.
Acknowledgement
The author would like to thank Ms. Chee Sok Jane for her research assistance
on this project. All errors in this paper, however, are solely my responsibility.
Endnotes
1. Compiled from various reports published by the Singapore Department of Statistics
(http://www.singstat.gov.sg
2. This is how the PECF report describes itself: Released on 15 April 2002, this report represents the
work of more than 150 individuals (comprising members of the various PEC Committees) from over
100 different organizations. It serves as a public education blueprint that complements the
Government’s initiatives in creating a total social environment conducive to marriage, families and the
raising of children. The Public Education Committee on Family was formed in September 2000 to
support the Ministerial Committee and Working Committee on Marriage and Procreation
(http://aboutfamilylife.org.sg)
3. The author is a team member of the Singapore-leg of the World Values Survey 2002.
4. Unless otherwise specified, all data presented are drawn from the WVS-Singapore 2002.
5. The so-called Fourth Wave of the WVS was conducted between 1999 and 2001 and covers 16
countries, some for the first time and others for the fourth time. Singapore’s data will be added to the
fourth wave. See http://www.worldvaluessurvey.org/ for more details on the WVS itself.
6. Statistically, the largest gender difference found in this survey was concerning the view: “Marriage
is an outdated institution”, for the subgroup of “divorced” respondents. However, this might be an
anomaly as there were very few respondents (9 male, 13 female), thus probably skewing the response
rate.
7. The age of the respondents for all the data reflects their age in the year 2002, when the World Values
Survey was conducted in Singapore.
Asia-Pacific Population Journal, April 2006
83
References
Alter, G. (1992). “Theories of fertility decline: A nonspecialist’s guide to the current debate”, in Gillis,
J. R., L.A. Levine, and D. Levine (eds.) The European Experience Of Declining Fertility,
1850-1970: The Quiet Revolution, Cambridge: Blackwell, pp. 13-30.
Beck, U. (1992). Risk Society, London: Sage.
Beck-Gernsheim, E. (2002). Reinventing the Family, Cambridge: Polity Press.
Coleman, J. W. (1998). Social Problems: A Brief Introduction, New York: Longman.
Gillis, J. R. (1996). A World of Their Own Making: Myth, Ritual and the Quest for Family Values, New
York: Basic Books.
Inglehart, R. (1997). Modernization and Postmodernization: Cultural, Economic, and Political
Change in 43 Societies, Princeton, N.J.: Princeton University Press.
Inglehart, R. and P. Norris (2003). Gender Equality and Cultural Change Around the World.
Cambridge, U.K.; New York: Cambridge University Press.
PECF (Public Education Committee on Family) (2002). “Family Matters!”, Report of the Public
Education Committee on Family (commissioned by the Ministry of Community, Youth and
Sports, Singapore), Public Education Committee on Family: Singapore,
(http://www.mcys.gov.sg/MCDSFiles/download/Family%20Matters%20(Full).pdf)
Quah, S. (1999). Study of the Singapore Family. Singapore: Ministry of Community Development.
________ (1998). Family in Singapore: Sociological perspectives (2nd ed), Singapore: Times
Academic Press.
Schwartz, S. H. (1994). “Are there universal aspects in the structure and content of human values?”,
Journal of Social Issues, No. 50, pp. 19-45.
Singapore Department of Statistics (2004) Report on Marriages and Divorces 2003. Singapore:
Singapore Department of Statistics.
________ (2002). “Twenty-Five Years of Below Replacement: Implications for Singapore”, paper
presented at Workshop on Fertility Decline, Below Replacement Fertility and the Family in
Asia: Prospects, Consequences and Policies - Twenty-Five Years of Below Replacement
Fertility, 10-12 April 2002, National University of Singapore, Singapore.
________ (2001). Singapore Census of Population 2000: Advance Data Release No. 8 (Marriage and
Fertility), Singapore: Singapore Department of Statistics.
Smart, C. and B. Shipman (2004). “Visions in monochrome: families, marriage and the
individualization thesis”, British Journal of Sociology, vol. 4, No..55, pp. 491-509.
UNDP (United Nations Development Programme) (2003). Human Development Report. Geneva and
New York: UNDP, 2003
Van Krieken, R. (1997). “Sociology and the reproductive self: demographic transitions and modernity”,
Sociology, vol. 3, No. 31, pp. 445-471.
World Bank (2002). World Development Report. New York: Oxford University Press.
Zimmerman, S. (2001). Family Policy: Constructed Solutions to Family Problems. Thousand Oaks,
California: Sage.
84
Asia-Pacific Population Journal, Vol. 21, No. 1
International Labour
Recruitment: Channelling
Bangladeshi Labour to East and
South-East Asia
Given that most host countries in East and South-East Asia lack
a viable alternative, dependence on migrant workers will be long term
even if they choose not to integrate them permanently into their
societies. Any migrant worker policy has to recognize that
such dependence is here to stay.
By Lian Kwen Fee and Md Mizanur Rahman*
International labour migration in Asia has experienced the most rapid growth in
the last few decades. There are two major destination regions for labour migrants in
Asia: Middle East and East and South-East Asia. In addition to countries of the
Middle East, since the early 1980s we observe a sustained growth of foreign
* Lian Kwen Fee, Associate Professor, Department of Sociology, National University of Singapore,
Singapore, e-mail: [email protected] and Md Mizanur Rahman, Post-doctoral Fellow, Asia Research
Institute, National University of Singapore, e-mail: [email protected].
Asia-Pacific Population Journal, April 2006
85
manpower in the prosperous countries of East and South-East Asia, particularly
Singapore, Malaysia, the Republic of Korea and Japan. Those countries have
followed specific temporary migrant worker programmes in recruiting foreign
workers although the name and nature of the programmes vary. One can identify two
types of temporary labour migration programmes implemented in the region – the
“work permit” and the “trainee” programmes. Each migrant worker programme
offers different rights and privileges to migrants. Malaysia and Singapore hire foreign
workers under the “work permit” system, which provides special benefits to them as
workers. But the Republic of Korea and Japan pursue a conservative policy with
regard to the admission of foreign workers. They hire foreign workers mainly under
the “trainee” system, which restricts benefits as trainees are not formally recognized
as workers. In general, labour migration policies in the receiving countries in Asia can
be broadly summarized as follows: limiting labour migration, limiting the duration of
migration and limiting integration (Piper, 2004: 75).
Recruitment constitutes an important part of the processes of this labour
migration. It is dominated by recruiting agencies and brokers, who act as
intermediaries between workers and foreign employers and those agencies and
brokers are responsible for the mobilization, recruitment, documentation and
transportation of workers overseas (see Jones and Pardthaisong, 1998; Skeldon,
1997). Whether it is an authorized or unauthorized form of labour migration, the
role of the migration institution is vital in the channelling of migrant workers,
without it few migrants would have the information or contacts needed for
successful migration. Given the reality of rapidly increasing numbers of recruiting
agencies in the home and host countries in the region,1 this paper examines the
perpetuation of labour migration through the prism of institutional theory.
Although some literature explicitly suggests that Asian labour migration is a
consequence of the rapid development of the “migration industry” (see Goss and
Lindquist, 2000; Skeldon, 1997), research has not adequately focused on the role
of institutions in contemporary labour migration. The authors attempt to advance
the knowledge by focusing on the recruitment of Bangladeshi migrant workers to
the countries of East and South-East Asia, particularly Singapore, Malaysia, the
Republic of Korea and Japan. Those four host countries have been selected
because they are the principal destinations for Bangladeshi migrant workers in this
part of the world.
While this paper largely discusses the role of migration institutions in
channelling Bangladeshi migrants under different migrant worker programmes, it
also focuses on the general consequences of such programmes and the policy
principles that can be pursued by these host countries to meet the demand for
86
Asia-Pacific Population Journal, Vol. 21, No. 1
migrant labour in order to minimize negative consequences. The paper is divided
into four sections: the first deals briefly with the significance of institutional theory
to labour migration; the second deals with various labour recruitment institutions
in Bangladesh; the third focuses on the role of different institutional actors in the
Bangladeshi labour migration to East and South-East Asia; and, finally, the
conclusion suggests policy recommendations.
Significance of institutional theory to labour migration
The greater part of migration research focuses on the causes for international
migration. To explain the causes of migration, macro-level theories address the
organization of socio-economic relations, the geographic division of labour and the
political mechanisms of power and domination (Portes and Walton, 1981; Amin,
1974; Wallerstein, 1974). Micro-level theories largely explain migration as a
consequence of either economic cost-benefit calculation of individuals or as a strategy
of households to diversify incomes and minimize risks – such as unemployment, loss
of income, or crop failures (Todaro, 1976; Lewis, 1954; Stark, 1991). However,
Massey and his associates (1994) argue that the conditions that cause migration may
be different from those that perpetuate migration. Two theoretical traditions have
been developed to explain the forces that perpetuate and sustain migration, network
and institutional theory. Researches have shown in great detail that migrants are
linked to each other through social networks and those networks are the social
infrastructure that sustains the migratory process (Massey, Alarcon, Durand and
Gonzalez, 1987; Boyd, 1989; Faist, 2000). Some authors also refer to networks, social
capital and individual human capital as the “engines of immigration” (Phillips and
Massey, 2000).
This paper argues that despite the contribution of the network analyses, the
investigation of perpetuation and sustenance of temporary labour migration flows
in Asia requires examination of the complex infrastructure of entrepreneurial
actors and activities that constitute the migration institution. Institutional theory
moves beyond individual or structural approaches to understanding the
perpetuation of international labour migration (Sobieszczyk, 2000: 393).
According to Goss and Lindquist (1995: 336), an international migrant institution
is usually a complex organization consisting of knowledgeable individuals and
agents of organizations (from migrant associations to multicultural corporations)
and other institutions (from kinship to the State). Individuals’ migration
experiences help to support migration institutions because successful migrants are
likely to make profit, become brokers of information and provide advice to others.
Institutional theory assumes that migrant recruitment policies and practices often
serve to create a black market in labour movement because of the high profits to be
Asia-Pacific Population Journal, April 2006
87
made by meeting the demand for overseas employment (Massey and others 1994;
1993). Some authors also refer to it as the “other engine of migration”
(Hernandez-Leon, 2005:2).
Some attempts have been made to focus on the role of institutions in
facilitating labour migration within East and South-East Asia. For example, Spaan
(1994, 1999) examines the specific role that middlemen and brokers play in
international migration. He provides insights into the role of different actors in
stimulating Javanese international migration to Singapore, Malaysia and the Arab
Gulf. In another, Sidney Jones (2000) discusses the various actors involved in
Indonesian labour migration to Malaysia. Jones and Pardthalsong (1998) in their
studies on Thai international labour migration offer an interesting explanation of
different migrant institutions in Thailand and their role in the migration process. In
another study, Sobieszczyk (2000) discusses in detail the development and
functions of migrant institutions in Thailand. Okunishi (1996) provides an
interesting analysis of different labour contracting systems in Japan. Wee and
Sim’s (2004) work focuses on the role of transnational networks in female labour
migration. They argue that this transnational network has functioned as a bridge
between the workers, especially female workers, and international labour markets.
Recently, Battistella and Asis (2003) examined unauthorized migration in
South-East Asia and provided an update on migration policies and paths within the
region. While those studies mainly deal with labour recruitment from and within
East and South-East Asia, there exists a dearth of scholarship on the recruitment of
migrant workers from any South Asian country to this region. This study attempts
to contribute to this under-researched area.
Labour recruitment system
Bangladesh
Bangladesh is a major emigrant country in Asia. According to one estimate,
the total cumulative figure for Bangladeshi migrants overseas until 2004 was
approximately 4 million and for East and South-East Asia alone it was around
450,0002 (table 1). Remittance has been a major source of foreign currency
earnings. Bangladesh received around US$ 32 billion as remittances from its
overseas migrant population between 1976 and 2004. Around 200,000
Bangladeshis annually migrated overseas for temporary employment in the 1990s.
At the end of 2001 a new Ministry, namely the Ministry of Expatriates’ Welfare
and Overseas Employment (MEWOE) was created to protect the interests and
rights of Bangladeshi migrant workers (Bruyn and Kuddus, 2005). Bangladeshi
migrants have started coming to the countries of East and South-East Asia mainly
after the Gulf crisis in 1990. In the migration process, prospective migrants
88
Asia-Pacific Population Journal, Vol. 21, No. 1
normally use services of five public and private institutions, which control the
functional linkages between Bangladeshi workers and employers overseas. They
are: (a) the BMET (Bureau of Manpower Employment and Training); (b) BOESL
(Bangladesh Overseas Employment Services Limited); (c) Recruiting agents; (d)
Sub-agents; and (e) finally, migrant-trafficking syndicates.
Two governmental bodies that fall under the MEWOE and deal with
international labour migration are BMET and BOESL. BMET was established in
1976 with the specific purpose of meeting the manpower requirement of the
country and for export of manpower overseas (Khondker, 2004; Rahman, 2003).
On the Government’s side, the BMET monitors and supervises the overall
recruitment process. It issues and renews the licenses of recruiting agencies, grants
permission to agencies to recruit, provides immigration clearances after verifying
visa papers and employment contracts. In 1984, the Government established the
BOESL as a limited company. The main purpose of creating this company was to
provide honest, efficient and quick services to valued foreign employers in the
matter of recruitment and deployment of manpower (see Siddiqui, 2001). The
agency coordinates with the Bangladeshi missions abroad in assessing the needs of
labour and puts up advertisements in newspapers for recruitment. M. Alimullah
Miyan (2003: 12) argues that BOESL operates as a model institution in the
manpower sector to work in healthy competition with the private agencies. The
BOESL usually deals with professionals and skilled migration.
Private recruiting agencies have come into being in the late 1970s to serve as
an important source of information for the growing prospective migrants. Those
agencies act as middlemen between foreign employers and local prospective
migrants. The recruiting agencies disseminate information through public
advertisement and prospective migrants contact them through brokers or
sub-agents to secure overseas employment. They charge the prospective migrant
workers for their services. After recruitment of workers, the list of recruits along
with signed contracts and other documents are submitted to BMET for clearance. A
sizeable number of individuals also manage to secure job contracts directly from
the employers overseas through personal networks. However, they are also obliged
to approach the authorized recruiting agencies to obtain clearance from BMET.
Thus, under certain circumstances the relationship between the migration industry
and migrant networks is one of complementarity. However, the role of recruiting
agents remains vital in the migration process. The recruiting agencies were
organized under the Bangladesh Association of International Recruiting Agencies
(BAIRA) in 1984 with representatives from 23 recruiting agencies. The
association had a membership of 475 agencies in 1998 and 700 in 2003. 3
Asia-Pacific Population Journal, April 2006
89
90
Asia-Pacific Population Journal, Vol. 21, No. 1
217
1,379
3,212
6,476
8,695
13,384
16,294
12,928
20,399
37,133
27,235
39,292
27,622
39,949
57,486
75,656
93,132
106,387
91,385
84,009
72,734
106,534
158,715
185,739
144,618
137,248
163,269
162,131
139,031
2,045,428
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
Total
643
1,315
2,243
2,298
3,687
5,464
7,244
10,283
5,627
7,384
10,286
9,559
6,524
12,404
5,957
28,574
34,377
26,407
14,912
17,492
21,042
21,126
25,444
22,400
594
5,341
15,769
26,722
41,108
397,687
Kuwait
1,989
5,819
7,512
5,069
4,847
6,418
6,863
6,615
5,185
8,336
8,790
9,953
13,437
15,184
8,307
8,583
12,975
15,810
15,051
14,686
23,812
54,719
38,796
32,344
34,034
16,252
25,462
37,346
47,012
499,429
1,221
2,262
1,303
1,383
1,455
2,268
6,252
7,556
2,726
4,751
4,847
5,889
7,390
8,462
7,672
3,772
3,251
2,441
624
71
112
1,873
6,806
5,611
1,433
223
552
94
1,268
93,895
Qatar
66,343
587
1,238
1,454
2,363
1,927
13,153
12,898
4,932
4,701
5,051
4,728
3,847
4,191
2,573
2,700
Iraq
173
718
2,394
1,969
2,976
4,162
2,071
2,209
3,386
1,514
3,111
2,271
2,759
1,609
471
1,124
1,617
1,800
1,864
1,106
1,966
1,934
1,254
1,744
1,010
450
1,574
2,855
606
52,937
335
870
762
827
1,351
1,392
2,037
2,473
2,300
2,965
2,597
2,055
3,268
4,830
4,563
3,480
5,804
5,396
4,233
3,004
3,759
5,010
7,014
4,639
4,637
4,371
5,421
7,482
9,194
107,524
Libyan
Arab
Jamahiriya Bahrain
113
1,492
2,877
3,777
4,745
7,352
8,248
11,110
10,448
9,218
6,255
440
2,219
15,429
13,980
23,087
25,825
15,866
6,470
20,949
8,691
5,985
4,779
4,045
5,258
4,561
3,854
4,029
4,435
236,239
Oman
17,237
4,921
85
28
224
258,040
2
401
1,385
1,628
10,537
67,938
47,826
35,174
66,631
2,844
551
530
23
3
23
1,558
3,315
2,759
889
578
1,501
990
1,561
28
3,771
215
17,166
229
776
642
313
1,739
391
3,762
5,304
27,401
21,728
9,596
11,095
9,615
6,856
5,304
6,948
116,296
110
385
1,083
331
178
718
792
25
Republic
Malaysia of Korea Singapore
228
328
1,335
2,659
3,062
303
169
1
1,420
2,958
154
980
1,802
15,412
Brunei
Darussalam
Table 1. Flow of Bangladeshi migration by country of employment
809
632
1,029
223
2
1,111
524
913
1,224
550
254
711
709
654
517
585
16
338
222
303
383
798
418
204
89
1,192
204
211
3,424
18,619
Others
6,087
15,725
22,809
24,495
30,073
55,787
62,762
59,220
56,714
77,694
68,658
74,017
68,121
101,724
103,814
147,131
188,124
244,508
186,326
187,543
211,714
231,077
267,667
268,182
222,686
188,965
225,256
254,190
272,958
3,924,027
Total
23.71
82.79
106.90
172.06
301.33
304.88
490.77
627.51
500.00
500.00
576.20
747.60
763.90
757.84
781.54
769.30
901.97
1,009.09
1,153.54
1,201.52
1,355.34
1,525.03
1,599.24
1,806.63
1,954.95
2,071.03
2,847.79
3,177.63
3,573.76
32,330.08
Remittances
(million US$)
Source: http://www.bmet.org.bd/Flow_Migration.htm, accessed in June 2005.
Note: Table excludes countries such as the United Kingdom, Lao People’s Democratic Republic, Mauritius, Jordan, Italy, Spain, Madagascar, Lebanon, Namibia and the category of
Miscellaneous provided by BMET for two reasons: they represent very insignificant numbers and this figure would not fit out in one page without excluding them.
Saudi
Arabia
Year
United
Arab
Emirates
Sub-agents form another level of institutional structure surrounding
migration. The term sub-agents refers here to those individuals who work in
collaboration with authorized or unauthorized recruiting agents as a link to
prospective migrants, recruitment agencies and overseas employers. Those
sub-agents are commonly known as brokers (dalals). There are mainly of two
types: local brokers and migrant brokers. Local brokers as the name suggests
are from the communities of origin of prospective migrants. Migrant brokers
are those enterprising migrants who are working overseas for a considerable
period. Because of their long migration experience, they acquire sufficient
knowledge about both authorized and unauthorized channels of migration, and
later use them for profits. The final group is the migrant-trafficking syndicate.
Migrant-trafficking syndicates involved in unauthorized migration are mainly
authorized and unauthorized travel agents. They are engaged in a range of
illegal activities like falsifying documents, changing the names and photos of
workers on documents, misusing visas and circumventing restrictions imposed
by receiving countries.
Singapore
Since the late 1980s foreign labour has had a significant influence in
Singapore’s economy. The main reason for relying on foreign manpower in
Singapore is relatively simple. Hui (2002) argues that the small size of the
domestic population could not have supported the rapid expansion of the
economy. Singapore’s economic growth between 1990 and 1998 averaged 7.9 per
cent per annum. He showed that, over this same period, more than 604,000 new
jobs were created against an increase of 458,000 in the domestic population and an
increase of about 300,000 in the domestic labour force. Clearly, without the
inflow of foreign manpower to supplement the domestic labour force, the
phenomenal growth in employment and GDP over that period would not have
been possible (Hui, 2002). Singapore pursues a transparent recruitment policy and
has changed it at times to meet the demand for foreign workers (see Wong, 1997).
The present work pass system provides four classes of foreign labour: Class P,
Class Q, Class S and Class R (for details see, Rahman and Lian, 2005). P-Passes
are for professionals, entrepreneurs and investors; Q-Pass for skilled workers;
S-Pass for technicians; and R-Passes for semi-skilled and unskilled workers
including domestic helpers. Presently, there are around 620,000 foreign workers
in Singapore (The Sunday Times, (Singapore) 13 November 2005). Of the
620,000 foreign workers, 540,000 are work permit holders and the remaining
80,000 are employment pass holders.
Singapore is a major receiving country for Bangladeshi migrant workers.
Presently, there might be as many as around 40,000 Bangladeshi migrant
workers.4 They predominantly come under the R-Pass category and are engaged
Asia-Pacific Population Journal, April 2006
91
mainly in the construction sector. Spaan’s figure was adopted to describe the
recruitment network of authorized migrants to Singapore (figure 1). As
Bangladeshi labour migration to Singapore is predominantly authorized
migration, recruiting agencies play a major role in the migration process. In
general, a prospective migrant visits the local broker to initiate migration. Those
local brokers work for both migrant brokers and recruiting agents from Dhaka.
Partial payment is a precondition for initiating the recruitment process. Once they
receive partial payment from prospective migrants, the brokers contact the
recruiting agents or migrant brokers depending on the prior agreement for job
placement. Once the recruiting agent or migrant broker receives the necessary
papers and partial fees from the local agents, they approach recruiting agents in
Singapore for an In-Principle Approval (IPA) for hiring foreign workers
(Khondker, 2004; Rahman, 2003).
Figure 1. Recruitment network of authorized migrant workers to Singapore
Fees
Fees
Prospective migrants
and family in
Bangladesh
Recruiting agents
(Authorized One)
Bangladesh
Brokers
(Local level)
Demand
Info
Info
Info
Supports
International border
Fees
Info and cash in case
of friends and
relatives
Migrant brokers in
Singapore
Info
Demand
Info
Migrants in
Singapore
Remittances
Recruiting agents,
Singapore
Migrant Flow
Transactions
Figure 1: Recruitment network of authorized migrant workers to Singapore
Note: The figure is adapted from Ernst Spaan, 1994
A prospective migrant requires an IPA to enter Singapore. This is issued by
Singapore’s Ministry of Manpower for a particular company on the basis of genuine
need for foreign workers. Recruiting agents in Singapore apply for IPAs. Because of
the limited number of jobs in Singapore and the huge supply of manpower in
Bangladesh, some local agents get involved in unhealthy practices, which expose
some prospective migrants to victimization. A first time unskilled migrant usually
pays between US$ 3,500 and 5,000 as fees for a two-year contract. However, for the
92
Asia-Pacific Population Journal, Vol. 21, No. 1
skilled and second time migration the fees vary noticeably (they can be as low as a few
hundred dollars). When a prospective migrant reaches Singapore, his employer
receives him at the airport and assumes all responsibilities including medical
examination, work permit, accommodation, and transportation to workplace. Upon
the completion of contract, it is the employers who are responsible for their
repatriation. Thus, low-skilled foreign manpower is managed as a temporary and
controlled phenomenon in Singapore (Yeoh, 2004: 19).
Malaysia
In the late 1970s, the international relocation of manufacturing industries
and the restructuring of the economy and society under the New Economic Policy
(1971-1990) created a large number of jobs for semi-skilled and unskilled
workers that many Malays were not willing to take. This generated a huge demand
for low skilled foreign workers. However, the Government of Malaysia only
acknowledged the need for foreign labour in the mid-1980s and the Cabinet
Committee on Foreign Workers implemented a five-year policy on
non-traditional source workers in 1991. Initially, the Government permitted the
hiring of foreign workers in plantation and construction. Later, it also allowed
hiring foreign workers in other industries like manufacturing and services
(Khondker, 2004). There were over 807,000 legally employed semi-skilled and
unskilled foreign workers in 2001 (Kassim, 2002). Malaysia is a recipient of a
large number of unauthorized migrant workers. Kassim (2002) reports that the
authorities have identified and apprehended over 2.25 million unauthorized
foreign workers under the various programmes in operations carried out between
1992 and 2001. To cope with such unauthorized migration, Malaysia passed the
Immigration Act 2002, which imposes severe penalty on irregular migrants (see
for details, Battistella and Asis, 2003).
Bangladesh was able to enter into an agreement with Malaysia for the
systematic transfer of labour in 1992 (Abul-Aziz, 2001). Following this
agreement, a large number of migrants started entering Malaysia through the
authorized channel. Bangladesh entered into another agreement with Malaysia in
1994 for the annual importation of 50,000 workers, mostly for the construction
industry (Ahmed, 1998). Athukorala and Manning (1999: 177) report that
307,000 Bangladeshi migrant workers were issued work permits between 1992
and 1998 for work in Malaysia. Bangladeshi migrants are mainly employed in the
plantation, manufacturing and construction industries. In 1997, Malaysia stopped
recruiting Bangladeshi workers following a huge influx of unauthorized
migration and amid reports of Bangladeshi workers marrying Malaysians.
Bangladesh and Malaysia signed another Memorandum of Understanding (MOU)
Asia-Pacific Population Journal, April 2006
93
in October 2003 concerning the hiring of Bangladeshis, which is yet to be
implemented.
In the 1990s, recruitment was carried out by private recruiting agencies and
brokers in Bangladesh and Malaysia. Malaysia had about 300 licensed recruiting
agencies to handle foreign worker intake (The New Straits Times, “Foreign Workers
must know Malay, English”, 12 November 2002). Authorized recruitment was
similar to Singapore; recruiting agents employed local brokers to identify the
prospective migrants and offered commission for their work. Local brokers usually
charged between US$ 2,000 and 3,000 for authorized migration and between US$
1,500 and 2,500 for unauthorized migration. Partial payment was obligatory to initiate
the migration process. The local brokers provided the necessary information and,
sometimes, financial assistance with higher interest rates to the prospective migrants.
Sometimes local brokers, without the consent from prospective migrants pursued
unauthorized means of migration for higher profits. Prospective migrants were
largely ignorant of the routes of migration – authorized or unauthorized at the initial
stage of their migration. They usually discovered their status once they were in
Malaysia. Even if they had known, they would have hardly challenged the
migrant-trafficking syndicates for fear of losing the fees paid in advance. Thus, it is
the local brokers who often determine the fate of prospective migrants and the means
of migration. The local brokers are usually rural elites and their power stems from the
traditional authority structure of villages in Bangladesh. They often escape from
prosecution owing to the lack of evidence of transactions between prospective
migrants and local brokers. After the financial crisis in 1997, labour migration to
Malaysia suddenly ceased, forcing aspiring migrants to seek the assistance of
migrant-trafficking syndicates. Those syndicates use two routes: one is through legal
visa procedure (usually applying for tourist, student or business visa) and the other,
unauthorized via second country, particularly Thailand (Ullah, 2006; Sabur, 1997).
Once in Malaysia, they contact migrant brokers or their relatives to help them find
employment and accommodation.
Republic of Korea
The Republic of Korea experienced a severe labour shortage of unskilled
workers in the small- and medium-sized industries in the late 1980s (Athukorala and
Manning, 1999). As a result, migrant workers came to fill the void in small
manufacturing companies. According to the Ministry of Justice of the Republic of
Korea, the number of migrant workers rapidly increased: 6,409 in 1987 and 245,399
in 1997 (OECD, 2002). Presently, there are around 421,000 foreign workers in the
Republic of Korea (The Korea Times, 3 March 2005). Principally, they are of two
types: migrant workers and industrial trainees. The first category comprises
94
Asia-Pacific Population Journal, Vol. 21, No. 1
professionals, technical workers and language teachers (Lee, 2002; Seol, 2000). The
second category comprises trainees. The Republic of Korea introduced a labour
import scheme known as the “Foreigners Industrial Training Programmeme (FITP)”
in 1991. This FITP was expanded in 1993. In 1994, the Government permitted the
Korean Federation of Small Business (KFSB) to recruit and manage prospective
trainees from 11 countries under 27 privately owned “international manpower
recruiting agencies” (Park, 1994). The manpower shortages were so serious that the
Programme beneficiaries had to be extended to include the coastal fisheries in 1996
and the construction industry in 1997 (Yoo, 2004).
Apart from those authorized foreign workers, there are a large number of
unauthorized foreign workers in the Republic of Korea. Unauthorized foreign
workers are usually those who over-stayed their visas or who changed their
sponsored employers in the case of trainees. The number of unauthorized migrant
workers was 188,000 in 2004 (The Korean Times, 3 March 2005). Yoo (2004)
reports that the Government of the Republic of Korea tried to convert the Industrial
Trainee Programme into the Guest Worker System since 1995, but failed to do so
because of opposition from business. However, since 2003, a public consensus has
grown in favour of the guest worker system for migrant workers. As a result, the
Act Concerning the Employment of Permit for Migrant Workers was enacted in
2003, and implemented in 2004 (see Yoo, 2004: 4). The guest worker system
however did not replace the Foreigner Industrial Trainee programme (FITP).
Along with the guest worker system, the FITP is also in operation. The
Government of the Republic of Korea has already signed a MOU with six Asian
countries namely Thailand, Viet Nam, Mongolia, Sri Lanka, Indonesia and the
Philippines for the deployment of some 25,000 workers in the country (The Korea
Times, 16 August 2004). Although Bangladeshi migrant workers have contributed
to the development of small- and medium-sized industries under the trainee
programme since its inception, they are deprived of serving as guest worker under
the recent guest worker programme for unknown reasons.
Bangladeshi workers began migrating to the Republic of Korea
immediately after the introduction of FITP in the early 1990s. According to
BMET, only 11,760 workers migrated there for work from 1994 to 2000
(Khondker, 2004). However, a Korean official source cites the cumulative
figure for Bangladeshi over-stayers in the peninsula between 1992 and 2000 as
69,600 (OECD 2000: 211). This means that a large number of migrant workers
took the unauthorized migration path to the Republic of Korea. For the purpose
of recruitment of Bangladeshi nationals, a few selected recruiting agencies
were permitted to send workers there. This created an opportunity to
monopolize recruitment by those selected agencies and to make huge profit off
Asia-Pacific Population Journal, April 2006
95
migrants. As the number of annual intake was limited to 5,000, many
prospective migrants became frustrated and looked for alternative channels.
Authorized migration is straightforward. Prospective migrants showed their
interest to the designated recruiting agents by paying service fees well in
advance. The designated recruiting agencies sent the necessary documents to
the Korean Federation of Small Business (KFSB) for the completion of the
recruitment procedures. Once the agencies received the processed papers from
the KFSB, they sent the prospective migrants by air. Although authorized
migration was supposed to be almost free of cost, migrants were charged
between US$ 2,500 and 3,500 in the early 1990s and between US$ 5,000 and
8,000 in the late 1990s. Currently, they charge up to around US$ 10,000 5 from
each migrant. Currently, unauthorized migration is the only alternative for
Bangladeshi migrants. Therefore, migrant-trafficking syndicates and migrant
brokers have come forward to filling the demand. Presently,
migrant-trafficking syndicates charge between US$ 10,000 to 14,000 for each
migrant. The payment of such large amounts of cash to migrant-trafficking
syndicates renders prospective migrants vulnerable to victimization. They are
willing to take the risk because the rewards are high if successful.
Japan
Japan’s economic success started attracting migrant workers from its
neighbouring countries in the 1980s (Nagayama, 1996). A large number of migrant
workers entered Japan during this period to join in the construction and
manufacturing sectors. Foreign workers in Japan can be divided into two groups:
authorized and unauthorized. Authorized foreign workers consist of Japan-born
Koreans and Chinese (Zainicbi Gaikokujin), foreign-born Japanese (Nikkeijin),
and trainees and entertainers from Asia (Iguchi, 2002). In 2004, foreigners
comprise 1.5 per cent of Japan’s population (Asian Migration News, 15 June
2004). While foreign-born Japanese are invited with long-term settlement
facilities, trainees are hired to meet the short-term manpower need. The Japan
International Training Cooperation Organization (JITCO), is responsible for
processing migrant trainees. The objective of the trainee programme is ostensively
to train workers from developing countries to learn skills in Japan that can be used
later back home for economic development. However, like in the Republic of
Korea, trainees in Japan hardly receive any “training”. Instead they work as regular
workers and receive allowances for their work.
There are a large number of unauthorized migrant workers in Japan.
Unauthorized migrant workers are mainly those who leave the sponsored
employers as trainees, overstay their visas (usually tourist, business or student
96
Asia-Pacific Population Journal, Vol. 21, No. 1
visas), or enter Japan with forged passports or by other unauthorized means.
According to data from the Immigration Bureau, the number of irregular migrants
in Japan, as of 1 January 2005, totaled 207,299 persons (Asian Migration News, 31
March, 2005). Thus, it is obvious that Japan has responded to the need for unskilled
labour by: (a) relying on ethnic Japanese from South America; (b) trainee scheme
to bring in “trainees” assigned to farms, forestry and fishery sectors, and small
companies and (c) turning a blind eye to more than 200,000 over-stayers who are
working in construction and other manual work. Scholars, such as Spencer (1992),
Nagayama (1992), Morita and Sassen (1994), Yamanaka (1993), Shimada (1994),
who work on migration issues in Japan concluded that irregular migration
stemmed from the tight immigration policy (see Battistella and Asis, 2003).
Bangladeshis began migrating to Japan in the early 1980s and it was
predominantly unauthorized migration. Mahmood (1994) estimates that 33,573
Bangladeshi entered Japan between 1985 and 1990. According to another estimate,
between 1990 and 1998, the cumulative figure for Bangladeshi over-stayers was
63,170 (Iguchi, 2002:127). However, many of those unauthorized migrants were
deported later. According to one estimate, around 5,078 Bangladeshi over-stayers
were deported between 1996 and 2000 (Kondo, 2002:427). As authorized labour
migration is not available for Japan, access to migrant-trafficking syndicates is
critical to successful migration. Migrant-trafficking syndicates usually use two
routes to send prospective migrants to Japan. First, they attempt to obtain a tourist,
student or business visa from Bangladesh. If the syndicates fail to get visas for
prospective migrants, they try for transit visa upon arrival at the international
airports in Japan. They usually use Thailand; Malaysia; Hong Kong, China; and
Singapore as springboards to enter Japan on arrival transit visa. Presently,
migrant-trafficking syndicates charge as much as $ 20,000 for each prospective
migrant. Prospective migrants may be abandoned at any point. Therefore,
clandestine migration to Japan is a risky venture. Failure means economic disaster
for the migrants and their families.
Conclusion and policy implications
By the 1980s the performance of the economies of East and South-East Asia
including the Republic of Korea, Japan, Singapore and Malaysia had reached a
level that further growth could only be sustained through the recruitment of
substantial migrant labour. Such workers were urgently needed to fill the void
created by the domestic population who regarded employment in construction,
manufacturing, shipping and plantation as dirty, difficult and dangerous. Both
Singapore and Malaysia, because of their colonial origins, have relied on the
contribution of migrants to economic development. The traditional sources of
Asia-Pacific Population Journal, April 2006
97
such labour were South Indian, South Chinese and Indonesian. By contrast, the
Republic of Korea and Japan have until recently drawn on domestic workers to
meet their labour needs. The two South-East Asian countries referred to in this
paper are essentially migrant societies with multi-ethnic populations. The two
East Asian societies are culturally homogenous and their exposure to ethnically
diverse migrant communities such as Bangladeshis is a recent experience. This
important difference is reflected in the policies adopted by their respective
Governments with regard to migrant labour. Singapore and Malaysia have more
open policies towards the recruitment of foreign labour migrants than either the
Republic of Korea or Japan.
Singapore has a rational foreign worker recruitment programme. It is finely
tuned and responsive to the needs of a changing labour market, which may range
from specialist and entrepreneurial skills to semi-skilled and unskilled workers in
construction and domestic services. It has implemented a work pass system that
grades potential migrants according to the levels of skill they possess. At the lowest
end, unskilled workers are given the opportunity to upgrade their skills and benefit
from their improved status, a provision that is not available in other countries.
Strict and effective enforcement of labour regulations together with the availability
of clear information to brokers and recruiting agents have contributed to a
transparent policy of migrant worker recruitment. The result is that Bangladeshi
workers are predominantly recruited through authorized channels.
Since its economy expanded rapidly in the late 1970s, foreign workers have
entered Malaysia illegally in significant numbers. This is partly owing to a
coastline that in many instances is only a boat ride for Indonesian, Thai and
Filipino migrants and partly to a liberal policy towards such workers and the
absence of a transparent recruitment policy. It was only in 1991 that it introduced a
policy of hiring migrants from non-traditional sources such as Bangladesh, but this
was discontinued after the financial crisis of 1997. By then, however, Bangladesh
migrant networks had been well established to circumvent the authorized channels
of migration. Formal recruitment was resumed in 2001 but in limited numbers after
lobbying from the plantation and construction industries. There is a mix of legal
and illegal workers in the country and brokers are familiar with utilizing both
authorized and unauthorized facilities to recruit workers. This simply reflects an
inconsistent labour policy as well as the lack of regular enforcement.
Both the Republic of Korea and Japan experienced a significant inflow of
migrant workers in the second half of the 1980s, as a consequence of severe labour
shortages in their small to medium size companies in construction and
manufacturing. Despite the urgent need for such workers in their economies,
98
Asia-Pacific Population Journal, Vol. 21, No. 1
longstanding public antipathy towards the presence of foreigners in those
societies has made it difficult for the State to introduce policies to regularize the
status of migrant labour. For this reason, Japan established a programme
supposedly to train foreigners to be sent to work in overseas Japanese companies.
The trainee programme was later adopted by the Republic of Korea. In reality it
was a backdoor way to facilitate the entry of migrant workers. The trainees
eventually leave the programmes to work illegally in industry for better wages.
While the Republic of Korea has a very limited scheme for the authorized
recruitment of migrants from non-traditional sources such as Bangladeshis, it
together with Japan relies predominantly on illegal workers. Publicly, the
Governments of the two countries have a stringent policy towards foreign migrant
labour. In practice, both adopt a liberal application of trainee recruitment and
visas to facilitate the use of illegal migrant workers. Under those circumstances
brokers have come to play an influential role in making available unauthorized
facilities for the entry of those workers.
In all the four countries surveyed it is clear that the recruitment of Bangladeshi
migrant workers, both authorized and unauthorized, is institutionalized within an
extensive network of agents, brokers and syndicates. The network extend from the
villages, districts and Dhaka of Bangladesh to the airports of the receiving countries;
and it provides a valuable source of information, contacts and support to facilitate the
entry of prospective migrants and their eventual settlement, however temporary. The
airports of those countries as well as others such as Bangkok, Jakarta and Manila also
serve as staging points from which Bangladeshis make numerous attempts to enter the
Republic of Korea and Japan. The effectiveness of this institutional complex is
appreciated if we take into account that specific districts in Bangladesh are identified
as particular sources of migrant labour for each of the four countries. The recruitment
system that has evolved over the years, even more so if they are unauthorized, has to
be understood as an integral part of the social organization of Bangladeshi migration –
encompassing household strategies, the ties of the extended family and the migrant
community in the host society. How well developed this organization is varies from
country to country and will hopefully be the subject of further research.
Table 2 summarizes the recruitment process of Bangladeshis in Singapore,
Malaysia, the Republic of Korea and Japan. Bangladeshi migrant workers,
whether documented or undocumented, are found in all receiving countries in
East and South-East Asia. They have used the available “temporary migrant
worker programmes” to enter as migrant workers or “trainees” except in Japan
where there is no legal programme for them, especially for those who are
semi-skilled or unskilled. Based on the recruitment experiences of Bangladeshi
migrant workers in those regions, “temporary migrant worker programmes”
Asia-Pacific Population Journal, April 2006
99
(including those disguised as trainee systems) are plagued with similar
consequences as Ruhs (2002) identified in North America, Europe and Asia.6
Namely, these are the emergence of illegal foreign workers; exposure of migrant
workers to various forms of exploitation; tendency of foreign workers to extend
the duration of stay; emergence of “immigrant jobs” in the receiving countries
(Piore, 1979); and finally, local workers’ opposition against the “foreign worker
employment programme”.
Temporary migrant workers programmes should be designed to serve the
interests of both sending and receiving countries. Devising systematic migrant
worker programmes that respect humanitarian considerations and recognize
genuine demand is timely. To address unauthorized migration, there is also an
urgent need to revisit some “migrant workers programmes”, especially in light of
criticisms of the treatment of legal migrant workers or trainees and the
programmes themselves. Given that most host countries in East and South-East
Asia lack a viable alternative, dependence on migrant workers will be long term
even if they choose not to integrate them permanently into their societies. Any
migrant worker policy has to recognize that such dependence is here to stay. Ruhs
(2002) advocates that all the stakeholders affected by this phenomenon should
have an input in policy-formation.
Castles (2000:12), drawing on Kassim’s work on Malaysia, comments “basing
policies on the assumption that the use of migrant workers is temporary, when they in
fact meet long-term structural needs, leads to non-compliance with official policies on
the part of both employers and immigrant workers. The result is that regulations are
often ignored and become unenforceable”. The proliferation of unauthorized
syndicates, and a network of agents, brokers and travel agencies involved in sending
Bangladeshi migrant workers to South-East and East Asia is a direct consequence of
this. For a start the trainee programme adopted by Japan and the Republic of Korea
should be replaced by a migrant worker programme attuned to the needs of those
economies. This is unlikely to happen in the short term given the political sensitivities
of the two Governments. Such a programme should accommodate workers from
unskilled to skilled, up to professional level; and provide opportunities for workers to
upgrade their skills through formal assessment and improved benefits. This
programme was introduced in Singapore in the late 1990s. Furthermore, foreign
workers who have been illegally employed for a minimum number of years should
either be allowed to exit with dignity or issued work permits since they have been
acculturated in the host society and have acquired the necessary skill and experience.
Those measures will contribute some way to minimizing the use of unauthorized
means of migrant labour recruitment and its abuses.
100
Asia-Pacific Population Journal, Vol. 21, No. 1
Asia-Pacific Population Journal, April 2006
101
Work permit
Work permit Department
Ministry of Manpower
Malaysians, Filipinos,
Bangladeshis, Indians,
Thais, Chinese, Indonesians
For first-time migrants,
between US$ 3,500 and
5,000. For second migration
(if skilled) below around $
1000.
Recruiting agencies
(around 1,000 agencies)
Authorized
Authorized
Strict
Air
Recruiting agents
Semi-skilled and unskilled
Responsible government
body
Major nationals recruited
Economic costs of
migration
Predominate role in
recruitment :
Dominant channel of
migration
Predominant type of
migrants
Attitude to irregular
migrants
Routes
Facilitators of
Recruitment
Skill Requirements
Singapore
Status given to migrant
worker
Items
Semi-skilled and unskilled
Recruiting agents &
migrant-trafficking syndicates
Air and land
Varied
Authorized and unauthorized
Authorized and unauthorized
Recruiting agencies
(around 300 agencies)
Documented migration:
Between US$ 2000 and 3,000;
Undocumented migration:
between US$ 1,500 and 2,500
Indonesians, Bangladeshis,
Thais, Filipinos, Indians
Pakistanis, Nepalese
Immigration Department
Ministry of Home Affairs
Work permit
Malaysia
Unskilled
Designated recruiting agents &
migrant-trafficking syndicates
Air
Liberal
Unauthorized
Authorized and unauthorized
KFSB (Korean Federation of
Small Business)
Documented migration:
around US$6,000 to 10,000;
Undocumented migration:
between US$ 10,000 and
14,000
Chinese (Korean origin),
Filipinos, Thais, Indonesians,
Bangladeshis, Pakistanis
KITCO under KFSB
Trainee
Republic of Korea
Unskilled: Not applicable for
Bangladeshis
Migrant-trafficking
syndicates
Air and sea
Liberal (Varied)
Unauthorized
Unauthorized
JITCO Not applicable for
Bangladeshis
Undocumented migration: as
much as US $ 20,000
Foreign-born Japanese,
Koreans, Filipinos, Chinese
JITCO:NA for Bangladeshis
Trainee: Not applicable for
Bangladeshis
Japan
Table 2. International labour recruitment in East and South-East Asia with reference to Bangladeshi migrants
Endnotes
1. For example, in Bangladesh there are currently around 800 registered recruiting agencies and around
3,600 travel agencies (1,667 authorized and 2,000 unauthorized) that are engaged in authorized and
unauthorized channeling of migrant workers (http://www.hrexport-baira.org/history_background.htm
accessed in January 2005). In the Philippines, there are around 2,960 government-registered recruiting
agencies
only
that
are
engaged
in
channeling
migrant
workers
overseas
(http://www.poea.gov.ph/cgi-bin/agList.asp?mode=all accessed in January 2005). In Singapore, there
were around 1,100 recruiting agencies in 1999 to serve the manpower need (The Straits Times, 24
October 1999, “Big Money in Labour Import Racket”).
2. This cumulative figure comes from the Bureau of Manpower Employment and Training (BMET), the
official source of Bangladesh. The BMET is responsible for keeping records of authorized migrant
workers but does not keep records for return migrants. Therefore, the authors assume that the actual
number of Bangladeshi migrants in this region will be higher than the BMET source suggests.
3. See BAIRA homepage, http://www.hrexport-baira.org/aims_objective.htm accessed in June 2005.
4. “Every year, more than 30,000 Bangladeshi workers … come here (Singapore)” (18 December
1999, “The Journey of Hope” The Straits Times). Usually, Bangladeshi migrants come on a two-year
contract. Therefore, one estimates that there have been around 40,000 Bangladeshi migrants in
Singapore in a given year since the mid-1990s.
5. The authors found several cases during their fieldwork in 2003 where prospective migrants paid
even more than US$ 10,000. The fees for migration to the Republic of Korea increased gradually over
time. Interviews reveal that in 2000 it was between US $ 6,000 and 8,000.
6. Ruhs (2002) examines the consequences of “temporary foreign worker programmes” in five
different countries (Germany, Kuwait, Singapore, Switzerland and the United States of America) and
finds: the emergence of “immigrant sectors” in the host country’s labour market; the vulnerability of
migrant workers to various forms of exploitation in recruitment and employment; the tendency of
temporary foreign worker programmes to become longer in duration and bigger in size than initially
envisaged; native workers’ opposition against the introduction or expansion of a temporary foreign
worker programme; and the emergence of illegal foreign workers who, together with local employers,
circumvent the programme.
102
Asia-Pacific Population Journal, Vol. 21, No. 1
References
Abdul-Aziz Abdul-Rashid (2001). “Bangladeshi migrant workers in Malaysia’s construction sector”,
Asia-Pacific Population Journal, vol.16, No. 1.
Ahmed, S. Nahar (1998). “The impact of the Asian crisis on migrant workers: Bangladesh
perspectives”, Asian and Pacific Migration Journal, vol.7, Nos. 2-3.
Amin, S. (1974). “Modern migrations in western Africa”, in S. Amin, ed., Modern Migrations in
Western Africa (London, Oxford University Press).
Asian Migration News www.scalabrini.org/~smc/amnews/amnarch.htm, by-weekly information
service hosted by the Scalabrini Migration Center, the Philippines, all issues (1997 to 2006)
are available online, accessed on 23 March 2006.
Athukorala, Prema-Chandra and Chris Manning (1999). Structural Change and International
Migration in East Asia: Adjustment to Labour Scarcity (London, Oxford University Press).
Battistella, Graziano and Maruja M.B. Asis, eds. (2003). Unauthorized Migration in South-East Asia
(Quezon City, Philippines, Scalabrini Migration Center).
Boyd, Monica (1989). “Family and personal networks in international migration: recent developments
and new agendas”, International Migration Review, vol. 23, No. 3.
Bruyn, Tom de and Umbareen Kuddus (2005). Dynamics of Remittance Utilization in Bangladesh,
IOM Migration Research Series, No. 18, pp. 1- 93.
Castles, Stephen (2000). “Migration as a factor in social transformation in East Asia”, Conference paper
on Migration and Development, Princeton University 4-6 May 2000.
Faist, Thomas (2000). The Volume and Dynamics of International Migration and Transnational Social
Space (Oxford, Claendon Press).
Goss, Jon and Bruce Lindquist (2000). “Placing movers: An overview of the Asian-Pacific migration
system”, The Contemporary Pacific, vol. 12, No. 2, pp. 385-414.
________ (1995). “Conceptualizing international labour migration: A structuration perspective,
International Migration Review, vol. 29, No. 2, (Summer, 1995), pp. 317-351.
Hernandez-Leon, Ruben (2005). “The migration industry in the Mexico-U.S. migratory system”,
Working Paper CCPR 049 (2005), California Center for Population Research, University of
California, Los Angeles.
Asia-Pacific Population Journal, April 2006
103
Hui, Weng-Tat (2002). “Foreign manpower policy in Singapore”, in Singapore Economy in the 21st
Century:Issues and Strategies, eds., Tee, Koh Ai, Lim Kim Lian, Hui Weng Tat, Bhanoji
Rao, and Chng Meng Kng (Singapore, McGraw Hill).
Iguchi, Yasushi (2002). “Foreign workers and labour migration policy in Japan”, in Yaw A. Debrah
(ed.), Migrant Workers in Pacific Asia (Great Britain, Frank Cass Publishers).
Jones, Sidney (2000). “Making money off migrants: the Indonesian exodus to Malaysia”, Hong Kong:
Asia 2000 Ltd. and University of Wollongong: Centre for Asia Pacific Social
Transformation Studies.
Jones, Huw and Teing Pardthaisong (1998). “The commodification of international migration: findings
from Thailand”, Tijdschrift voor Economische en Sociale Geografie, vol. 90, No. 1, pp.
32-46.
Kassim, Azizah (2002). “Economic slowdown and its impact on cross-national migration and policy on
alien employment in Malaysia”, Migration and the Labour Market in Asia: Recent Trends
and Policies, Organisation for Economic Co-Operation and Development, pp. 325-339.
Khondker, Habibul Haque (2004). “New trends and changing landscape of Bangladeshi migration”, in
International Migration from South Asia, ed., Hisaya Oda (Japan, Institute of Developing
Economies.
Kondo, Atsushi (2002). “The development of immigration policy in Japan”, Asian and Pacific
Migration Journal, vol. 11, No. 4, pp. 415-436.
Lee, June J. H. (2002). A Review of Data on Trafficking in the Republic of Korea, Prepared for IOM,
Geneva, International Organization for Migration.
Lewis, W. Arthur (1954). “Economic development with unlimited supplies of labour”, Manchester
School of economic and Social Studies, vol. 22, pp. 139-191.
Mahmood, Raisul Awal (1994). “Adaptation to a new world: Experience of Bangladeshis in Japan”,
International Migration, vol. 32, No. 4, p. 194.
Massey, Douglas S. and others (1993). “Theories of international migration: a review and appraisal”,
Population and Development Review, vol. 19, No. 3, September.
________ (1994). “An evaluation of international migration theory: The north Mexican case”,
Population and Development Review, vol. 20, No. 4, December.
104
Asia-Pacific Population Journal, Vol. 21, No. 1
Massey, Douglas S., Rafael Alarcon, Jorge Durand, and Humberto Gonzalez (1987). Return to Aztlan:
The Social Process of International Migration from Western Mexico, University of
California Press.
Miyan, M. Alimullah (2003). “Dynamics of labour migration – Bangladesh context”, Conference
Paper,
IUBAT,
December
31,
2003
http://www.kli.re.kr/iira2004/pro/papers/MAlimullahMiyan.pdf accessed March 2005.
Morita, Kiriro and Saskia Sassen (1994). “The new illegal immigration in Japan 1980-1992”,
International Migration Review, vol. 28, No. 1, pp. 153-163.
Nagayama, Toshikazu (1996). “Foreign workers recruiting policies in Japan”, Asian and Pacific
Migration Journal, vol. 5, Nos. 2-3, pp. 241-264.
________ (1992). “Clandestine migrant workers in Japan,” Asian and the Pacific Migration Journal,
vol. 1, Nos. 3-4, pp. 623-636.
OECD (2002). Migration and the Labour Market in Asia: Recent Trends and Policies, Organization for
Economic Cooperation and Development.
________ (2000). Migration and the Labour Market in Asia.
Okunishi Yoshio (1996). “Labour contracting in international migration: the Japanese case and
implications for Asia”, Asian and Pacific Migration Journal, vol. 5, Nos. 2-3, pp. 219-240.
Phillips, Julie and Douglas S. Massey (2000). “Engines of immigration: stocks of human and social
capital in Mexico”, Social Science Quarterly, vol. 81, pp. 33-48.
Piore, Michael J. (1979). Birds of Passage: Migrant Labour and Industrial Societies (Cambridge:
Cambridge University Press).
Piper, Nicola (2004). “Rights of foreign workers and the politics of migration in South-East and East
Asia”, International Migration, vol. 42, No. 5, pp. 71-95.
Portes, Alejandro and John Walton (1981). Labour, Class and the International System (New York,
Academic Press).
Rahman, Md Mizanur (2003). “Bangladeshi Workers in Singapore: A Sociological Study of Temporary
labor Migration”, Ph.D. Dissertation, Department of Sociology, National University of
Singapore.
Rahman, Md Mizanur and Lian Kwen Fee (2005). “Bangladeshi workers in Singapore: the view from
inside”, Asia-Pacific Population Journal, forthcoming.
Asia-Pacific Population Journal, April 2006
105
Ruhs, Martin (2002). Temporary Foreign Worker Programmes: Policies, Adverse Consequences, and
the Need to Make them Work, Working Paper No. 56, June, 2002, The Center for
Comparative Immigration Studies, University of California-San Diego.
Sabur, M. A. (1997). “The plight of Bangladeshi migrant workers”, Asian Migrant, vol. 10, No. 2,
April-June, pp. 64-65.
Seol, Dong-Hoon (2000). “Foreign workers in Korea: Issues and discussions”, Philippine Journal of
Third World Studies, vol. 15, No. 1, pp. 115-138.
Shimada, Haruo (1994). Japan’s “Guest Workers”: Issues and Public Policies (Tokyo: University of
Tokyo Press).
Siddiqui, Tasneem (2001). Transcending Boundaries: Labour Migration of Women from Bangladesh
(Dhaka, University Press Limited).
Skeldon, Ronald (1997). “Migration and Development: A Global Perspective” (Harlow, Longman).
Sobieszczyk, Teresa (2000). “Pathways abroad: Gender and international migration recruitment
choices in northern Thailand”, Asian and Pacific Migration Journal, vol. 9, No. 4, pp.
391-428.
Spaan, Ernst (1994). “Taikongs and calos: the role of middlemen and brokers in Javanese international
migration”, International Migration Review, vol. 28, pp. 93-113.
________ (1999). Labour Circulation and Socio-economic Transformation: The Case of East Java,
Indonesia, Report No. 56, Netherlands Interdisciplinary Demographic Institute (NIDI),
The Hague.
Stark, Oded (1991). The Migration of Labour (Oxford, Basil Blackwell).
Spencer, Steven A. (1992). “Illegal migrant labourers in Japan: Coping with a new invasion,”
International Migration Review, vol. 26, No. 3, pp.754-786.
Todaro, Michael P. (1976). International Migration in Developing Countries: A Review of Theory,
Evidence, Methodology and Research Priorities. A WEP Study, Geneva: International
Labour Office.
Ullah, AKM Ahsan (2006). “Bangladeshi labour migrants in Malaysia: the experience on the routes”,
unpublished Conference Paper, Association for Cultural Studies Crossroads Conference,
20-23 July 2006, Istanbul Bilgi University.
Wallerstein, Immanuel (1974). The Modern World System I: Capitalist Agriculture and the Origins of
the European World Economy in the Sixteenth Century (New York, Academic Press).
106
Asia-Pacific Population Journal, Vol. 21, No. 1
Wee, Vivienne and Amy Sim (2004). “Transnational networks in female labour migration” in Aris
Ananta and Evi Nurvidya Arifin, eds., International Migration in South-East Asia, Institute
of South-East Asian Studies: Singapore.
Wong, Diana (1997). “Transience and settlement: Singapore’s foreign labour policy”, Asian and
Pacific Migration Journal, vol.6, No. 2.
Yamanaka, Keiko (1993). “New immigration policy and unskilled foreign workers in Japan,” Pacific
Affairs, vol. 66, No. 1, pp. 72-90.
Yeoh, Brenda S.A. (2004). “Migration, international labour and multicultural policies in Singapore”,
Working Paper Series, No. 19, Asia Research Institute, National University of Singapore.
Yoo, Kil-Sang (2004). “Migrant workers’ labour market in Korea”, Conference Paper, 5th Asian
Regional Congress of the International Industrial Relations Association (IIRA) held in
Seoul on June 23-26, 2004, convened by the Korea Labour Institute (KLI) and the Korea
Industrial Relations Association (KIRA).
Asia-Pacific Population Journal, April 2006
107
Asia-Pacific Population Journal Guidelines for contributors
The Asia-Pacific Population
Journal is a periodical produced three
times a year by the Emerging Social
Issues Division of the Economic and
Social Commission for Asia and the
Pacific (ESCAP), with support from
ESCAP and the United Nations
Population Fund (UNFPA). Its purpose
is to provide a medium for the
international exchange of knowledge,
experience,
ideas,
technical
information and data on all aspects of
the field of population in order to assist
developing countries in the region in
improving the utilization of data and
information for policy and programme
purposes, among others.
Original
contributions
are
invited, especially papers by authors
from or familiar with the Asian and
Pacific region. Ideally, those papers
will discuss the policy and/or
programme implications of population
issues and solutions to problems and
report on experiences from which
others may benefit.
paper should not exceed 7,000
words, including tables, figures,
references, etc. It should include a
short abstract (100-200 words) of
the issues addressed and the most
important findings.
A complete list of references
arranged alphabetically by author
should also be included at the end
of the manuscript. Please refer to
examples in any issue of the
Journal. Figures and tables should
be supplied on diskette, preferably
in any major spreadsheet program
or in Microsoft Excel.
Manuscripts are accepted on
the understanding that they are
subject to editorial revision.
Contributors should indicate in an
accompanying statement or letter
that the material has not previously
been published or submitted for
publication elsewhere.
Note to contributors
A brief introduction about the
author(s) (title and affiliations, and
so forth) should also be included.
All material submitted for the
consideration of the Editorial Board
should be in the English language. The
paper, prepared in one of the major
word-processing programs, should be
supplied in “hard-copy” form with a
diskette (3 1/2 inch) or sent by e-mail
attachment to the address indicated
below. Ideally, the printed copy will be
typewritten in double spacing on one
side only of white A4 paper. The
margins should be generous, at least 3
cm (roughly 1 inch) wide, preferably
more for the left-hand margin. The
As all manuscripts will be
subject to peer review by
professionals in the field, the
name(s) of the author(s) or other
identifying information should be
placed on the title page only in order
to preserve anonymity. Manuscripts
may be sent by airmail or e-mail to
the Editor of the Journal, Emerging
Social Issues Division, ESCAP,
United
Nations
Building,
Rajadamnern
Nok
Avenue,
Bangkok 10200, Thailand, e-mail:
[email protected]
CONTENTS
Viewpoint:
Will HIV/AIDS Levels in Asia Reach the Level of Sub-Saharan Africa?
By John C. Caldwell
Potential for Reducing Child and Maternal Mortality through
Reproductive and Child Health Intervention Programmes: An Illustrative
Case Study from India
By Minja Kim Choe and Jiajian Chen
Readiness, Willingness and Ability to Use Contraception in Bangladesh
By Haider Rashid Mannan and Roderic Beaujot
Singapore’s Family Values: Do They Explain Low Fertility?
By Alexius Pereira
International Labour Recruitment: Channelling Bangladeshi Labour
to East and South-East Asia
By Lian Kwen Fee and Md Mizanur Rahman
UNITED NATIONS ESCAP
United Nations
Economic and Social Commission for Asia and the Pacific
Emerging Social Issues Division
United Nations Building, Rajadamnern Nok Avenue
Bangkok 10200, Thailand
Tel:
(66 2) 288-1586
Fax:
(66 2) 288-1009
Web site: http://www.unescap.org
E-mail:
[email protected]
[email protected]
Vol. 21, No. 1 April 2006
Readers of the Journal are invited to suggest topics for future coverage
and to comment on the articles already published herein.
Correspondence should be addressed to the Editor, Population and
Social Integration Section, Emerging Social Issues Division.
ASIA-PACIFIC POPULATION JOURNAL
Articles: